Skip to main content


See other formats





soc::flL HISTORY OF ::NDI 



M.A. CCal.X D. Phil. tOxon. 1 ) : Diplomee in Education, Oxford. Reader 

sn Education, Patna University, 1925-26, Sometime Vice-Principal, 

Training College, Patna, and Ofa Principal, Training College, Cuttack. 

Professor of Indian History. Patna College, Patna. 



- - - ' 

KBffl *WJBlwaattifKfei<-) M HV^ T1aj ^ 

''">n-,l, m, 

*' jj 


'' Vr' ', 


London, Edinburgh, Glasgow, Copenhagen, New York, Toronto, 

Melbourne, Cape Town, Bombay, Calcutta, Madras, Shanghai, 

Humphrey MHford, Publisher to the University. 




( V ) 


This book is the Thesis on " Some Aspects of the Earliest 
Social History of India," by which Dr. Subimal Chandra 
Sarkar gained the degree of D. Phil. here. It is the outcome 
of extensive research, not only in the Vedic and other brah- 
manic literature, but also in tlie Epics and Puranas. He has 
dealt with it in a fresh manner., independent of preconceived 
ideas and accepted views, and has brought together a great 
quantity of new evidence regarding the, social conditions of 
ancient India, that has been hitherto neglected, presenting it 
generally in new connections and a new light. There can be 
no doubt that the Mahabharata and the older Puranas, which 
are largely secular literature., disclose many real features of 
the ancient society that cannot be discarded or belittled, though 
they find no place in the priestly literature and differ from the 
brahmanic presentment, for in any case the existence and 
preservation of such different notices must be accounted for. 

^ One inference_that such independent research appears to 
elicit is that the Aryans, when they entered India, found in 
places a degree of civilization as high as their own, if not 
higher, especially in Oudh and North Bihar; and there need 
be nothing surprising in that, because it has happened more 
than once in tlie history of the world that a more virile tribe 
has overcome and entered into a higher civilization, and has 
afterwards carried that on to further excellence. 

This book is therefore well worth study, and should help 
to revise views that may now be held on insufficient grounds. 

OXFORD : December, 1924, 

i vii ) 



On my way to Nalapda, in September 1923, I spent two 
pleasant clays at Patna, Pataliputra of Buddhist fame, now 
on of the principal seats of learning in India, under the 
hospitable roof of Dr. S. G. Sarkar. We had many an 
interesting conversation on problems of Indian literature, and 
amongst other things he showed me the Manuscript of his 
Doctor's Dissertation on the Earliest Social History of India. 
The subject was of the greatest interest to me. Glancing over 
it I could see that it touched on some subjects which I had 
myself dealt with several times during the last thirty years, 
the first time in my paper on Ancient Indian Marriage Ritual 
in 1892, and the last time in my essays on Woman in Brah- 
mardsni (1920). There was no time to read the dissertation 
then and there. But Dr. Sarkar kindly gave me a type-written 
copy of it that I might read it at leisure during my voyage 
home. This was made impossible by .a prolonged illness 
which befell me after the completion of my happy pilgrimage 
to and through India, even before 'I reached the shores of Italy. 
Thus it was not until Easter 1924 that I could read the 
dissertation. Now I read it with delighted interest, though in 
many details I could not agree with the author, and I read it 
even more than once, in order to re-examine his arguments, 
where I differed from him. But from the very beginning I 
highly appreciated the scholarly instinct with which he has 
extracted from the Yedic texts every little detail that had even 
the least bearing on social life. 

Thus, in the first chapter, on Building of Houses, etc., 
he is not content with arranging all the passages referring to 
architectural details,- but he collects at the same time every- 
thing that can in any way elucidate the economical conditions, 
and the social and political condition of Ancient India. In the 
chapter on Household Furniture, and again in that on 
and Costumes, we find many references to marriage customs and 
married life, and even to ethnical and racial distinctions. Here 
he touches, for instance, the vexed question of the Vratyas, 
whom he takes to be Easterners and " non-Ailas " (non- 
Aryans) , adopting the terminology of Pargiter. In a paper 
on the Yratyas that has just been published (in the Zeitschrift 
fur Buddhism VI, 1924-25, p. 48 ff.\ I have, like 
Dr. Sarkar, also come to the -conclusion that the Yratyas were 
neither wandering Sadluis nor Saiva mendicants, as 'some 
scholars have tried to, prove, but certain tribes, living outside 
the pale of Brahraanism, and that there are some indications 
of their having been Easterners. I do not think, however, that 

( viii ) 

It Is possible to decide whether they were Aryans or uon- 

But the most interesting chapters of the dissertation are 
doubtless those on Sex-relations and the Status of Women in 
Ancient India. There are many things in these chapters to 
which I would take exception. 

Thus I certainly should not conclude 1 from the Yotlic 
myths that the Rsis of old did not see anything wrong in such 
connexions as that of Prajapati with his daughter, or of J.Tisan 
with his mother and sister. Surely the -ancient (i reeks did 
not approve of fathers eating up their children, because accord- 
ing to the Greek myth Eronos devoured his children. 1 am uro 
Dr. Sarkar himself would not believe that the Ksi \\iio said 
that Agni, as soon as born, eats his mothers ' or parents 
(jayamano niatara garblio atti : Kv. X. 79, 4), approved of 
children eating up their parents. 2 

[l I am glad that the learned professor has raised points for it 
would serve ^ to illustrate how it is sometimes difficult even for 
deep and critical scholarship to completely overcome tho subtlo 
influences of ancient prejudices and traditional or preconceived 
interpretations. I hope however that the footnotes 1 have 
ventured to add here may lead to a subsequent modification of 
the views of a scholar in whose soundness and fairness of judg 
ment I. have a very great faith indeed. AUTHOK.] 

t2 It will be, noted that my conclusion is not based on any out; Vodic 
myth or two; and one of these so-called myths (viz., that of 
Prajapati and his daughter) I have shown to be a brahmaiucal 
version of a secular dynastic detail. The basis of my inferences 
is not only these two references to Pusan's or Prajapati's 
conduct, but a number of other more disLiact allusions 
in priestly as well as secular historical literature. In- 
cestuous connexions and cannibalism are not analogous 
or parallel features in the history of civilization; the former 
may i_i^ discc > ve red even in comparatively recent history us an 
established feature, while the latter, so far as the history" of tho 
more civilised races is concerned, can only be inferred from 1'aiut 
echoes in folklore and myths. It cannot however bo denied that 
some ancient Hellenic traditions and myths are echoes of a 
remote period of barbarism, witchcraft, human sacrifices, and 
perhaps even of cannibalism. A scientific historian is surely 
justified in surmising from the Vedic (or rather pro- Vodic) 
Agni legends, not that the Vedic rsis were cannibals, but that 
these are relics of a forgotten barbarous ago, when the Indian 
tribes amongst whom fire worship arose (and I have shown 
tnem to have been pre-Aryan and Gangetic) still retained racial 
memories of the well-known primitive practice of ontiii" up the 
old members of the tribe either after (samficial) shunter or 
exposure and death. So also it is very likely that the auciout 
Hellenes found traditions of buch a primitive practice Interim' 
* m ^ gS l earlie - r Mediterranean people, which quite naturally 
found their way into the mixed Greek mythology. Finally it 
will also be remembered that parental incests wore not unknown 
amongst .ancient Greeks and Persians, whose cultural' affinities 
w th ancient Lido-Aryans are clear enough. In investigating 
all such details we should steer clear of the perfectly natural 


The, stories told in the late Jataka commentary, not In the 
old Jataka gathas, about Eama and Sita, cannot prove that 
Sita was common wife of Eama and Laksmana, nor that Sita 
was their sister as well as wife. Generally speaking, though 
the existence of incestuous marriages must be admitted for 
Ancient India, as it is found among other ancient peoples, I 
do not believe that it existed to such an extent as it would 
appear from the statements made in this dissertation. In my 
opinion it never was, even in primitive times, a- general popular 
custom, but limited to ruling families or dynasties. 1 

Nor can I approve of the author's explanations of the 
Gandharva in the wedding mantra : ' ' Somali prathamo vivide 
Gandharvo vivida uttarah, trillyo Agnis te* patis turlyas te 
manusyajah." The exact nature of the Gaudharvas Is certainly 
still one of the unsolved problems of Indian mythology and 
folklore. Still it is clear enough that Soma, the Moon, is 
considered as the ' husband * of the maiden on account of 'his 
regulating the menstruation, and that Agni was called the 
1 husband J of the bride from whom the mortal husband receiv- 
ed her, on -account of the time-honoured custom of leading the 
bride around the fire at the wedding ritual. In the same* way 
the Gandharva. Visvavasu must somehow be related to the 
sexual life of woman (the Buddhist Assalayana Suit a, shows 
that he was connected with conception; see also Ev. X. S5 ? 

U Here again, Sita/s consanguinity and biandry (or potentia* polyandry) 
has not been inferred from only one reference in the Jatakas 
but t also from many other corroborative allusions there as well 
as m JLpic-Puranic literature, taken together with contemporary 
Vedic evidence on the subject. Occurrence of incestuous mar- 
riages "among other ancient peoples" is nut however the 
soundest reason for inferring their existence in ancient India. 
1 hough of course this has its confirmatory value; it is the first- 
hand evidence of the priestly and secular historical literatures 
that I have relied upon. " The statements made in the dis- 
sertation " are not fanciful, and references have been given for 
all statements, which will have -to be taken for what they are 
worth irrespective of the attractiveness or otherwise of the 
conclusion. It rather puzzles me that while the equally late 
and much tampered with Kavya version of the really ancient 
Eaniayanic traditions is passed by scholars, the Jataka or Bud- 
dhist version, which from Ihe standpoint of historical criticism 
is a much sounder source, should be viewed with - unmitigated 
scepticism. I have not jumped to a conclusion that consan- 
guinous marriages and polyandry were " general popular 
customs"; I have only suggested that the evidence available 
points to a frequent occurrence amounting to a custom of such 
connexions amongst the chief ruling as well as priestly families 
of the Vedic (=Epic) age. ATJTHOR.] 

f., and Av. IV, 37 fj, and had certainly nothing to do 
with' the higher education of girls. 1 

But I must not enter into further details. The book will 
doubtless meet with sharp criticism and arouse .strong opposi- 
tion. Some of the conclusions arrived at by the author will be 
accepted as true, others will have to be rejected. But errors 
are not only unavoidable, they are more often than not even 
necessary stages on the way to the discovery of truth, if only 
the search after truth is carried on in a truly scientific spirit. 2 
And even the opponents will admit that this is the case in 
Dr. Sarkar's dissertation. 

Though we may hesitate to ascribe to the traditional 
genealogies and legends of the Puranas so much historical value 
as our author, a faithful disciple of Mr. Pargiter, ascribes to 
them ; yet as an historian he is fully justified in trying to find 
out what light the Epic and PurSnic traditions might throw 
on the history of the Vedic period. In our days, when some 
scholars hold that there is no real tradition at all connecting 
the hymns of the Kgveda, which are believed to have been 
composed somewhere in Iran, if not still farther "\\ r est ? with 

fl Here tlie only difference between Dr. Wiiitornifcz and myself is that 
he takes Gandaarva to be connected with the soAual lil'e of 
Cornell, -while I take it to !e connected with some pre-iparital 
part of woman's life. The Gsndharva Visviivasii is certainly of 
a sexual character, but he is also a ! Muse *; besides Visvavasu 
is not named in the mantra in question. Thai Gandhavva is not 
always a- sex-spirit is shown by Tedic references Lo ' gandharva- 
griiita * maidens and lady-teachers. There i.s no real conflict. 
between the two interpretations, for the sexual character of 
spirits is very closely related in ancient (or even modern) thought 
with their artistic character. Soma's connection with menstru- 
ation would apply equally well, perhaps better, to my view of 
this wedding mantra : this interpretation of Soma's significance 
would make the education oF girls in music and arts begin -with 
adolescence. the most- suitable age for it; moral discipline or 
ritual purification (lepresented by Agni's ' husband-dom ') would 
naturally come after it, leading to real and perfected wifehood- 
Agni can hardly have been regarded as a husband of the bride 
simply because the marriage ceremony included going round tho 
fre; the fire was only the divine witness; from tho * sex ' point 
of view the stone, on which the Vedic bride mnnnlod for the 
sake of progeny, would be a more suitable candidate for the 
husband status in the mantra. Agni is very prominently con- 
nected with the ' brahmacarya ' of boys; why then nol of girls, 
\vlio, as the Av. says, could gel properly married only by 
passing through ' brahmacarya ' or a, period of education of some 
sort ! If ' gandbarva * in the wedding mantra is taken to b a 
* conception s spirit, then the absurd result would follow that 
"Vedic society credited every bride with one or more previous 
conceptions "before being led to the fire-'altar, unless it can bo 
sliown that this particular mantra (in isolation from tho rest) 
was originally intended for legalising illegitimate connexions 
^ with issue thereof, Agni's function being ' s'uddhi.' ATITHOB.] 
2 With t!ie*p remarks I entirely agree. - 

the later Indian literature, it is worth something to have shown 
that there are after all some threads that lead from the Egveda 
to the Brahmanas, and from these to the Epics and the 

Dr. Sarkar, has derived from the Puranas many startling 
facts and suggestions, specially as regards the sexual morality 
of the highest classes of society in ancient times. How far 
the suggestions will stand the test of criticism and become 
* facts ' remains to be .seen. I -am myself rather sceptical 
about some of these suggestions ; yet I cannot help admiring the 
absolutely unprejudiced and truly historical spirit in which the 
whole investigation is carried on by the author. And therefore 
I have great pleasure in recommending the book to all scholars 
who are interested in the historv of Ancient India. 


It only remains for me to express the hope that Dr. Sarkar 
may not be prevented by his official duties from devoting 
himself to scholarly work and continuing" the researches which 
be has BO happily begun. 

PRAGUE : November 9, 1924. 




SECTION A. p agea 

Settled life house propertyterms for the dwelling-house; 
their uses -significance cf some house-names derived 

from features of construction ... ... ... i 5 

Types of building-construction associated with pastoral 
and agricultural life with social life (domestic and 
public) with political life with educational life ... 513 

Roads causeways bridges fords and rest-houses ... 1315 

Building activities connected with Kings arid lesser 1527 


Kings, nobles and vassals : cas'Lles the forts and 

defences of the Dasas of the Aryans ... ... 15 22 

Fortified towns and capital cities gateways ... 22 24 

Royal establishments palaces, etc. ... ... 24 27 

Constructive activities in the village ... ... 27 28 

Features and plan of the Atharva-vedic dwelling-house, 
(hay, reeds, bamboo, clay, bricks, etc. ; wings, veran- 
dahs, etc.) 

Identical with the Lower Gangetic " bungalow " style 
representing an Eastern, Deltaic, and Brahmanical 
style of architecture ancient, indigenous, and with 
Dravidian affinities ... ... ... ... 28 32 

Indications of a style of timber architecture 

Characterized by doorways, pillars, etc. associated use 
of metals in house-construction -a Middle-Himalayan 
and submontane style, with Aryan and Mongoloid 
affinities ... ... ... ... ... 32 35 

Indications of a structural style characterized by use of 
burnt bricks 

altar-construction varieties of bricks cement altar 
models and sizes tradition re invention of bricks 
(Istaka) the brick- style an advanced and ancient 
one traditionally Brahmanical, Lower Gangetic, and 
with Dravidian affinities ... ... ... 35 38 

Indications of a stone architectural style 

Eastern (heterodox) and Midland (orthodox) styles of 
' smasana ' construction the Round and Stone style 
the Square and Brick style prototypes of the 
Buddhist (Stupa) and Brahmanical architectures the 
Hound Stone style belongs to* Magadha-Praci great 
antiquity of the* * smasana ' structures of R/ound^ and 
Square types and of the 3 varieties of ' smasana ' 
structures : reliquaries (round or square), memorial 
buildings, memorial pillars the prototypes of Bud- 
dhistic structures : ' stupas, 9 dedicatory buildings, 
moijplithic pillars the true explanation of the Mauryan 
stone-architecture the stone style is Magadhan, non- 
Brahmanical, proto-Buddhistic, and very ancient ... 38 46 


SECTION A. (CW</.) Pages, 


Three well-marked structural styles in the V-edic Age, 
corresponding to clear ethnic and regional divisions, 
and agreeing w-th facts of traditional history ... 45 


j ETC. " 47 _ 55 

Xature of references to furniture : ritualistic, brtilimanical 
and indigenous ... ... ... "... 47 

Primitive types of seats in ritual ... ... ... 47 

Here advanced types of seats riparian materials the 
products of indigenous industries of the Lower Gangetic 
countr y ...... . .- ... 48-49 

Bedsteads, couches, and ' secular * sea'is : of timber, etc. 
sometimes modified with riparian materials in conformity 
TOh ritual tradition ... ... .,. _ 4950 

Varieties of bed-steads, couches and settees chiefly con- 
nected with women ... ... ... J ... 50 _ 52 

The aristocratic seats ; ' asandl ' and ' paryanka ' 

I V^tra^and Eastern origin-adopted by Midland rulinc 
| families-ritual modifications summary of descrip- 
tions, bringing out the general type of these seats, their 
different uses, modifications., dimensions and materials 52- 56 


W Variety m materials and manners, due to different 
regional conditions and tribal peculiarities H6rC _ 


Varietiesr-their use is bralunanical. Vrabya 



"' - ... 57-58 


stuff not ritual ancieit anrf tf T i l woollen 

aa Norlh-WeVwn /rJL^ i f. n v mt Et ' c ular, latoi 

t^^B^^&g ^e of *S" c * ld i : id ^ stri91 

BuDstiintPs of * woollers M ^fi r,n 

'-it ' oo-'-uu 




SECTION G.(Contd.) . 


Fully deveiop-ed and long-standing indigenous industry- 
very prominent in earliest brahmanical texts applian- 
ces and products same as those of the well-known 
Gangetic coUon manufacturers (Dravidian affinities) the 
6 vasas ' and its parts, described and identified (with 
the traditional * dhuti ' and ' sari ') similar varieties 
of borders, fringes, colours, etc. 

Manner of wearing the c vasas ' 

The 'various styles -of doing the "nivi" the true 
meaning of ' nivi ' the Vratya stj-le women's style 
later parallels 

Upper garments, over-garments, etc., of men and women : 
Varieties and uses of scarfs, veils, cloaks, tight jackets, 
bust-bodices, etc., described and identified * atka,' 
'pesas' and ' 6amulya,' specially brahmanical and in- 
digenous adoptively Indo-Iranian 

The turban 

Non-brahmanical Vratya. origin used by ruling' classes 
ritual modifications brahmanical head-bands 
3 head-bands non-JLrvan affinities ... 



Early use in martial equipment only use and materials 
of shoes in ritual apparently iion-brahmanical (ex- 
cluded by foot-ornament K) specially used by Vra'tyas 

Prevalent fashions of hair-diessing 

Naturally excluding Kurba-ns a specially hrahmauicaJ 
feature brahman clans (and indigenous gods) witE 
distinctive styles of hair-dressing the various styles 
affected by women styles common to both sexes those 
specially feminine descriptions Dravidian affinities ... 









Evidence of priestly literature -re 

(i) Traces of primitive sex-relations and special customs ... 
(a) Brother-sister marriage 

(6) Parental incest 

(c) Polyandry 

(d) c Niyoga ' (levirate) 

(e) Widow-burning and Widow-remarriage ... 

(f) Polygamy 







SECTION A. (Contd.) 

(ii) features of the normal marriage-forms 

Child-marriage unknown stages -of its subsequent 


Evidence re marriageable age and free love-matches bet. 
grown up parties 

Opportunities for pre-marital lovessocial freedom 
Extent of parental control ; limits and exceptional cases 

A Teat variety of normal forms of marriage tribal 
" customs, etc., the so-called ' Vedic ' marriage does 
not describe any one type 

The variety also dne to manners of settling a match, and 
exceptional conditions ... 

Marriages not restricted 1 to specified groups consanguinity 
no bar subsequent restrictions ' intercaste ' marriages 
frequent hypergarny not the main form ... 

(Hi) Social position and relations of women ... 

Significance of Vedic 'terms denoting woman indicative 
of her status and relations 

Attitude towards female b'rths no infanticide 

Relation between daughters and parents : honour, appre- 
ciation, concern for marriage, domestic work, leisure 
and liberty for social enjoyments, consideration for old 
maids, property and dowry, ' rich ' daughters, better 

legal position 

Personal and social freedom of the daughter initiative 
in love affairs gome fraternal control social equality 
in sex-relations 

Admission of women to highe&t education indications of 
regular systems of schooling for girls unmarried 

-women teachers 

The allegory of the ' life-stages ' of a woman ever in a 
state of marriage 









1091 1 



Evidence of * bardic tradition* re primitive sex-relations 

and special customs ,,. ... ... ^ 113225 

(a) Introductory- 
General agreement with evidence of priestly lit. ^pla* 
nation of the nature of traditional evidence, its valffe 

m tn j*l 4- 9n iw^t^r^** . ^.i . ^-. ^1 .. ^ *_.^ i i* i . * 

1 1 ~ ***^ wj. ui w; vi 1LH \Jliai C V lU.t511t"Cj HiS 

and the proper standpoint from which to view it 



SECTION B. (Contd.) 

(b) P.uranic IllustrationSj and critical estimates of the two 
sets oi evidences re 

(t) Brother-sister marriage 
(ii) Parental incest 
{Hi} Polyandry 
(iv) ' Niyoga ' (levirate) 

(r) Widow-remarriage an^i Widow-burning 
) Polygamy 





(Houses, etc.) 

THE Vedic Aryans very early ceased to be wandering 
tribes : the Rgveda shows them indeed still conquering, but 
they have already begun internecine wars and struggles for 
overlordships 1 ; and fighting does not involve constant shiftings 
of abode. Permanent settlements, of the nature of marks 3 2 
are normal in the Egveda, being the ' ksitir dhruva ' 3 or the 
fixed secure abode of the clans ; such were the " vis ' (in its 
special sense) , 4 the ' stha ' 5 (inferable in the early ' gostha, 3& 
or the later ' sthapati '), 7 the c vrjanas ' 8 and the ' vrajas.' 9 
It^is a settled life that could give the home its appropriate 
epithet of ' pratistha ' 10 or establishment, standing, fixed abode ; 
so also, one desiring 'to lead a settled home-life of his own is 
called " pratistha-kama.' 11 Such a settled home is already the 
nucleus round which the Yedic society and polity develops. 
Religion, law and custom was thus based on home-life, and the 
individualistic tendency of the Indo-Aryan found expression 
in, and grew put of the* importance he attached to' the home. 

Already in the early Yedic times, bouses were not simply 
unit family abodes, but were also individual private properties, 
which could be .acquired 12 ; and sometimes a ready-made house 
could be purchased for a 'Considerable price 13 fa well-to-do 
person possessed several houses ; thus a rich householder is 
called * pastyavant,' 1 * and some poet-singers are described as 
' puru-dama ' 1S ; so also (later on) fields and ' ayatanas ' 1& are 
given a-s examples of prosperity. 

The great variety of Vedic words denoting a dwelling- 
house is a reflection of its importance to the Vedic Indians, and 
shows that they were long settled, with a tradition of house- 
building. Gaya ' is a common word 17 in the Egveda for the 
house or household, inclusive of the inmates -and their belong- 

1 As the Puranic tradition amply shows, 

2 Cf. Roth: Diet., s.v. * vrjana/ 

3 Cf. Rv. I, 73, 4; VII, 88, 7; Zimmer: Alt. Leb. 142. 

4 E.g. where ' grha ' is contrasted with ' vis ' : Rv, X, 91, 2; cf. VII, 

op 1 

5 Cf. the Germanic ' Stadt.' 
o V^IQB inirfl 

? Cf. Kat. gr. But. I, 1, 12; Weber: Bid. Stud. 10, 13, 

8 Rv. I. 51, 15; 73, 2.4; 91, 21 5 105, 19; 128, 7; 165, 15; 166, 14; etc.; 

VII, 32, 27; X, 42, 10; etc. 

9 Rv. X, 179, 2=Av.' VII, 72, 2. 

10 Av, VI, 32, 3=VIII, 8, 21 = 23ankh. Aran. XIL 14. (Zimmer: Alt. 

Leb. 181, sees in it a legal term, but cf. St. Pet. Diet.). 

11 Taitt. Sam. II, 1, 3.4; Pane. Bra. XXIII, 1S : 1; etc. 

12 Cf . ' vida'tha * ; also n. 4, p. 4. 

13 Av. IX, 3, 15. 

14 Rv. I, 151, 2; IX, 97. 18 (prob.). 

15 Av. VH, 73, 1. 

16 Chand. Upati. VII. 24, 2. 

17 Rv. I, 74, 2; V, 10, 3; 44, 7; VI, 2, 8, etc.; Av. VI, 3, 3 Vaj. Sam. 

XXVII, 3. 

{ 2 ) 

ings; so is ' dama ?1 (and ' dam ') meaning house or home, 
implying an idea of control, 2 or possibly of building 3 ; 
'dMinan,' 4 another word for a dwelling or house, also signi- 
fies on the one hand l the inmates of the house,' 5 and on the 
other ' law or ordinance,' 6 showing the connexion in the Veclic 
mind between the house and all conceptions of law and order. 7 
' Sanna 18 is a house as a comfortable place, ' mahi ' (big) 
and * smat 5 (fine), within the ' vis ' or ' vrjana.' ^Grha,' 9 
the family home is contrasted 10 with ' jana ' and ' vis,' just 
as the family sacrifice is contrasted 11 with the sacrifice of the 
4 jana ' or c vis/ the individualism of the home being clearly 
recognized. * Ivuk ' in the compound * kulapa ' (used of the 
house-protector or family-chief, 12 and the home-staying 13 old 
maid) conveyed the sense of the dwelling-house o'f a small 
individual family, a sense which also occurs in the post-Vedic 
use 14 of the word singly : though later on the word acquired 
an added special meaning of ' sanctuary or temple/ 15 ' Vasati ' 16 
and k nivesana, ' 17 seem on the other hand to have been terms 
without special significance ; the former probably remained so 
all along, 13 but the latter is used in the Epic and the Puranas 
in the sense of a flourishing or fresh ' colonial settlement/ 19 
and in the Sutras in a curious optional sense of ' resting-} >Iar,e 
or stall for cattle' as opposed to the k grha ' used by men. 20 

1 Rv. I, 1. 8; 61. 9; 75, 5; 143, 4; II, 1.2; ftc. ; Vaj. Sum. VllT, 24 

2 Cf. Roth : St. Pet. Diet, s.v- ' dama. 5 ' 

3 Cf. V.I, 1, 540, s.v. ' dama.' 

4 Rv. I, 144, 1; II, 3, 2; IE, 55,10; VIII, 61, 4; 87, 2; X, 13, 1: 

etc.; Av. IV, 25, 7; VII, 68, 1; XII, 1, 52; Vaj. Sam. ],V, 34: 
Taitt. Sam. II, 7. 2. ' 

5 Bv. VIII, I0i, 6; IX, 36 ; 14; X, 82, 3; Av, II, 14, 6; (cf. St. Pet. 

Diet, s.v., c,}. 

:-IV 5 > 2; VI ' 21 > 3; VI1 ' 63 ' 5 5 Vni 41 ^ 10 ^ ^ 48,11. 
hich also comes out in the Egv-edic expression ' rta-dhaman *' (Rv T 

123, 9; IV, 7, 7; VII, 36," 5; X, 124, 3). ' " ' 

8 Rv. VII, 82, 1 ; I, 51, 15. J 

9 See also infra, for other uses of this term 

10 Rv. X, 91, -a. 

11 Bv. VII, 82, 1. 

12 Rv. X, 179, 2. 

IS Av. I, 14, 3; etc. 

14 Sat Bra. I 1 2 22; H, 1, 4 4; 4, 1, 14; XI, 5, 3, 11; 8, 1, 3; XIII, 

15 E* irM 5 B f i d ' J U P^ ( 1 ' 5 ' 32; Ohand, Upan, III, 13 6, etc. 

ur ^uia,' But cf ' ' > 

cognate g ur ^uia,' But cf. ' kula vad'hu > and 

IS Rv.l 15; V2, jVaj. Saip. XVIE, 15;Taitt. Bra. H, 3, 5, 4; 

iXA t , Oj o ; Gic. 

1? Rv- IV 19, 9; VII, 19. 5; (sanse of colonial settlement, 


ma no, ^- ' 7asati ' is eiven the 

mart or trading settlement or quarte? of a town Of 

etc. Cf. a 


Pastya ' (f) 1 or ' pastya' (n), 2 occurring singly, or in the 
compounds ' pastya -vant,' 3 ' pastya-vant >3 and l pastya-sad, M 
are other terms denoting a house or dwelling, and hence family, 
while in the feminine iorui even the goddess of tlie homestead 
may be so designated. 5 ' Pastya ' was occasionally also applied 
to the k stall for horses, ' the whole being u^ed for ft part, e.g., 
in ' asvii-pastya ' 6 and ''pastya-vant icarya' 7 ; but it had 
usually, along with ' hararya," 8 a special significance of * the 
home with all its adjuncts and surroundings/ 4 the family 
settlement,' apparently a nobleman's nbode di^vinj' f tab 'CM, 
etc.). l Yastu ' seems to mean simply " dwelling-house ' or 
k settlements generally ?9 in the compound epithet l su-vastu ' 10 ; 
but in ' va-stos-pati Jl1 it approaches the later (even modern) 
and more special meaning of ' the site of a house ' 12 ; these 
imports of k a group of houses ' or ' settlement/ and of * a site 
presided over by some deity/ are also conveyed by ' pastya * 
in several passages. 13 * Mana ' is a house as being a measured 
structure, wherein the house-builder saw a spirit ' manasya 
patni/ mistress of the hoiise-struetnre. 14 s Ayatana-/ 
4 enclosure/ had an earlier general sense of ' abode ' or 
f home/ 15 bat later on was specialized in use, like ' knla/ and 
referred to some sacred .structure within such enclosure. 15 
' Vis ' is a term which gradually narrowed in significance, 
from ' .settlement, ' 16 to the assembly-hall of the settlement/ 
and then to ' any house/ as is shown bv the uses of the 

1 Rv. I, 25, 10; 40, 7; 164, 30; IV, 1, 11; VI, 49, 9; VII, 97 ; 5; 

IX } 65, 23: X, 46, 6; (also corresponding passages in Yv. ). 

2 Rv. X, 96, 10.11; cf. VIII, 39, 8; VI, 53. 2; IX. 98, 12; V, 50. 4. 

3 Rv. I, 151, 2; IX, 97, 18; II 5 11, 16; IV, 54, 5; VTII ; 7, 29; (IV, 

55, 3; VIII, 27, 5). 

4 Rv. VL 51, 9; Roth. St. Pet. Dic. 5 s.v. ; Pischel : Ved. Stud. 2, 211. 

5 Rv. IV', 55, 3; VTIX 27. 5. 

6 Rv. IX, 86, 41 : Av. VI. 77, 1 : XIX, 55, 1. 

7 Rv. IX, 97, 18; prob. I, 91, 13. 

8 Cf. VJ. 3 I, 229, '30; Rv. VII, 56, 16; cf. X, 106. 5. 

9 Cf. th$ similar use of vei-n. ' vasti ' (from 'vasali'). 

10 Rv. VIII, 19, 17; (Nir. IV, 15). 

11 E.g. in Taitt. Sara. Ill, 4, 10; cf. Macdonell : Ved. Myth,.. 138; 

Zimmer: Alt. Leb., 236. 

12 \s opposed to the ' ksetra,' holding, also presided over by a deity; 

cf. Rv. IV. 37, 1 2; etc., Av. IL 12. 1; etc. 

13 E.o-. in Rv. VIII. 7, 29; VIII, 27, 5; IV, 55, 3: respectively; cf. 

Pischel's explanation of e pastya '=a river, having groups of 
houses on its banks. . 

14 Av. IX, 3; III, 12; cf, the later structural technical terms: 

" ' vastumaiia ' (in Pur.), * mana-sara ' (the treatise). 

15 E.g. in Ohand. Upan. VII, 24, 2: so alsn in the Epic ; it is applied 

subsequent! v to temples and monasteries enclosed by walls. 

16 Rv TV. 4, 3; 37, 1: Vf 3, 5; VI, 21, 4; 48, 8; VII, 56, 22; 61.. 3; 

70, 3; 104, 18; X, 91, 2; etc. (But in gome of these the sense 
'* dwelling-house ' may also suit.) 

( 4 } 

compounds ' vis-patf' 1 and ' vM-patmV* A cognate term 
' vesman ' 3 denotes ' house as the place where one is settled.' 
House or holding In its aspect of acquired property, 4 is desig- 
nated by ' vidatha ' ; but its specialized derivative meanings 
are quite early and manifold, amongst which may be noted 
those of ' asylum,' 5 ' family assembly or sacrifice/ 5 ' a 
smaller 7 or secular assembly/ 8 ' a rich or royal establishment 
like palaces.' 9 

A few common house-names were derived from ordinary 
features of building construction, such as ' ayatana/ 10 referring 
to the enclosing walls, railings or fencings; or ' durona/ 11 
4 gateway/ secondarily implying a house as characterized by 
such a feature. Another structural term is ' sala/ primarily a 
thatch of ' paddy-straw/ 12 for shelter of men or their cattle 
and stores, then the ; homestead inclusive of such stalls and 
sheds/ 13 and finally 'house' generally, as in ' salapati/ 
4 ho use -holder/ 14 or even a section or a single room of a house,, 
as in ' patni-sala ' 15 or ' agni-sala ' 15 ; apparently it came, to be 
quite early used of flourishing and wealthy residences as jvell, 

1 Rv, I, 12, 2; 26, 7; 164, 1; II, 1. 8; III, 2, 10; 40, 3; VII, 39, 2; 

IX, 108, 10; X, 4, 4; 135, 1; etc.; cf. also VIII, 55, 5=Av. IV, 
5, 6. 

2 Taitt. Sam. in, 1, 11, 4. 

3 Kv. X, 107, 10; 146,3; Av. V, 17, 13; IX, 6, 30; Ait. Bra. VIII, 

24 6 

4 J. Am, Or. S., 19, 12flj cf. Rv. I ; 117, 25; II, 1, 6; X, 85, 26.27; 

Av. XVm, 3, 70. - 

5 Ludwig: Trans. Rv. 3, 261; cf. Sat. Bra. V, 3, 1, 13, arid Kat. 

gr. Siit. XV, 3, 35; cf. also Rv. I, 31, 6; V, 62, 6; Ait. Bra. 
I, 30, 27.28. 

6 According to Bloomfield and V.I. ; cf. connexion of women chiefly 

with this, but rarely with the * sabha ' ; cf . also Kv. X, 85, 26.27 \ 
Av. VII, 38, 4; Mait. Sam, IV, 7, 4. 
-1 Zimmer: Alt. Leb., 177; Ev. II," 27, 12. 

8 According to Both; Rv. II. 1, 4; 27, 1217; III, 38, 5.6; V, 63, 2; 

VH, 66, 10; VEX 39,' 1 ? X, 12, 7; Av. I, 13, 4; XVII, 3, 15. 

9 Rv. IV, 27, 2; cf. I, 91, 20; 167, 3; Av. XX, "128, 1, 
10 Vide E. 15, p. 3. 

n Vide p. 32, n. 4 5, and p, 33, n, 14; 'dur/ 'durya' and *duryona* 
also hav<e a similar secondary sense. 

12 As * sala * is a term practically confined to the Av. (vide infra.), it 
is highly probable that it represents some indigenous word, 
presumably the same as the Eastern vern. t cala,' of equivalent 
form, and of exactly the same significances ( ' cala * and * cal ' 
also having a common figurative sense of house, room, etc.) ; 
K vern. * cal ' (rice) corresponds to Sans. ' sali ' ; cf . ' vicali ' = 
straw, i.e. * taken out of sali or paddy plants ' for E. vern. 
'c^Sans. * s,' cf, infra., l kasipu * = Tamil ' kacci-pa." 

B Av. m, 12, lfi s V, 31, 5j VI, 106, 3; VIH, 6, 10; IX, 3, Iff; 
XIV,, 1, 63; Taitt, Bra, I } 2, 3, 1; Sat. Brr. HI, 1, 1, 6; etc. 

14 Av. IX, 3, 12. 

15 Vide pp. 30-31. 

( 5 ) 

as indicated by names of princes and noted priests, like 'Maha- 
sala ' ('big-housed') l Pracisa-sala ' (' ancient-housed *)\ 

It is significant that ' sala/ etc., do not occur even once 
in the BY., while almost all the references to them belong to 
the AY., which applieslhis term also to a particular type of 
* straw and bamboo ' house 2 whose construction it describes. 
On the other hand ' durona/ etc. /are specially Bgvedic terms, 
while ' ayatana ' belongs to the Upanisads and the Epic. In 
the comparatively drier and hotter Upper G-angetic regions, 
the ' -entrance ' and ' enclosure ' aspects 3 of the dwelling-house 
must have been naturally more prominent 4 than the protective 
covering overhead : and the references to these features and 
their figurative use, accordingly, occur in texts that. were 
mainly of Midlandic origin ; -again, it is only in the rain-flooded 
Lower Gangetic country that the roof is naturally all-important, 
and has to be built carefully 5 : and accordingly, the ' sala ' 
(thatch) is prominent, and means the house itself, in tests 
that were largely of Eastern Gangetic (and indigenous) origin. 
So also in subsequent developments of Indian architecture, 6 
these two main styles may be recognized : one characterized 
by various modifications of the ' enclosure ' 7 and the s gate- 
way', 8 another by those in roofing 9 ; and it is remarkable that 
c roof ' architecture throughout the greater part of India (and 
in all periods) bears a distinct stamp of the Lower Gangetic 
* cala ' (sala), whether we look to the dome of the l stupa/ 10 
the convex ' saiva ' n roof with projecting eaves, or the curved 
and tapering ' vaisiiava ' n ' sikhara.' 

It is quite in accordance with the pre-eminently agri- 
cultural^ and pastoral character of early Yedic life tliat the 
house is at first very closely associated with the stalls for 
domestic animals. Thus the cowstall, the wagon and the 

1 Eg. Sat. Bra. X, 3, 3, 1 ; 6, 1, 1 5 Cha-nd. Upan. V, 11, 1 ; Muiid- 

Upan. I, 1, 3 ; cf. the early royal naineb * Maha-sala ' and 
' Vi-sala ' in the Puranic dynastic lists. 

2 Vide infra., p. 28ff. 

3 Cf. the sense of c enclosure ' in ' vrjana ' and * vrajV which is 

also described as ' sargala ' and * saparisraya ' (with gate and 
palisade); vide infra.; theso terms also are * specially Rgvedic. 

4 As it is even to-day. 

5 As the modern P,W.D.' knows very well, 

6 E.g. in Mauryan and post-Mauryan examples. 

7 E.g. the timber palisades or stone- railings. 

8 E.g. the famous * torana,' a form comparable to ' durona/ which 

may have been the prototype, an ornamental gateway, instead 
of an ordinary * dvar (a),' 

9 E.g. the so-called c barrel-shaped ' tops of monasteries, etc. 
JO In Buddhist i.e. Magadhan styles._ 

11 Miscalled ' Dravidian ' and ' In do- Aryan ' respectively by Fergusson; 
really they are both developments from the same Bengal thatch 
or t cala,' adapted to local conditions (vide Havcll's works for 
proper interpretation). 

house are mentioned toy ether 1 in the same breath as it. were. 
* Sala ' 2 and k pastya(a) ' 5 imply .accommodation of some isoii 
for both men and their beasts. 4 ' Gotra ' and ' vra('a)jn./ all 
originally arrangements for accommodation of cattle, were so 
intimately connected with the ordinary life of their possessors, 
that these names came to be employed equally or almost at 
the same time with reference to men. Thus ' viaja/ 5 pen, 
also denotes a pastoral .settlement (under a chief) 6 'including 
many k kulas ' and L vrajas ' un the narrower ; hi sub- 
sequent literature also (classical and modern), 'vraja T (pos- 
sibly also the representative of the older 'vrjana T has the 
regular sense of a closely organized pastoral' settlement with 
the human and bovine elements equally prominent. 8 What 
the ' vraja ' originally was, does not clearly appear ; (Jeklner 
uenvus A irom ' u-aj,' to go, giving it the primary mecimng 
oi pasture, while Both prefers the derivation from ' vrj '' 9 
\\indi gives the primary meaning of enclosure or pen - prob- 
ably both -senses are mixed up in the passages where it occurs 
the later (vernacular) use of ' vraja ' 10 agrees with this view 
the frequently occurring sense of pen 11 or stall" carmoi , )e 
denyed from vraj,' to go, but the sense of ' pasture 3JL3 is 
possible trom vrj ' to enclose; for a common pasture may 
*erv well have been an enclosure with a hedge- fence or pali- 
sade; it seems that such a defensible 'enclosure ' with mli- 

a \ v\ j-1 j-v 1 f^ -^ *~* *. j fj < *t i^ ^"'f'vVIAlo 

I i a I iL/ 1 4 I f^T T7 OS "T "^ J. i *\ll ~\ *4- t> L j^ *. i *-^l 4 1 y I 

Si oS^f SSl^m!^^ ' >^- ^'hc 

os T ve 

'>' ia oS f s it k ;," 

2 Cf. n. 13," 'p. 4. 

T /""fJ? f "" * 

Specially horses' in fh e latter case. 

7 Cf. n. 8, p. 1 
3 P! n- ; 

^rna episodes. 

j .^ ~~ - - ' r * ^J t L T , V^. * 

. probably Rv X 

Probably fa ! Ev/n .J T 8 ' 12 ' 2; V ^ Sai - I, 25 

nWels of the Jaina and ^"S...^ . ?? ave . on the 

w ^ - .- - - -- - -T rtir*^l* f -*4.A\J, 

K Taitt. 

17 ? L v. XT"? J ^ J "" i<5 ' a; (cf - ^T*a. Sam'"! 2^ 

_^ * ' * i * * ^ 1 Uw J 

2 v - IT - 51 ' 2 - 

is Kans. Bra. H, 9. 

( 7 ) 

to have the primary meaning of ' herd/ 1 which alone he thinks 
would explain its later use as 'family' or 'clan.' 2 But 
Iloth's interpretation of it as k cowstall ' as a structure 3 is 
better : firstly, as the suffix k tra ' is also indicative, of place ; 
secondlv, as the sense of a whole clan can easily be derived 

*J *"" 

from the sense oi a cowstall, common and spacious, wiiere a 
whole clan kept their cattle 4 ; and thirdly, as k gostlia ' 5 * i* 
similarly used of the Bharata clan, and ' gosthi ' 5 later on, by 
a similar transition, comes to mean a social circle. Geldner 
thinks that in all passages where ' gostha ' occurs, 6 the yense 
of ' grazing ground ' is better and suits all. 7 But here again, 
Whitney's and BloomfielcTs rendering of stall or stable 8 is more 
appropriate, as the ' stha ' points to some sort of a standing 
structure, a stand or stall, and cannot, evidently, refer to 

* grazmg ' : so that 'gotsha ' would mean liteiaily the standing 
place for cows. It is significant that even in modern verna- 
cular ' gotha ' is alwavs contrasted with ' matha ' (meadow) , 

dj W t- * 

with which it is combined to form a phrase. The use of 

* gotsha ' in Ait. Bra. 9 is interesting : the cows of the Bharatas 
are there said to be in the ' gostha ' at evening and In the 
' samgavim ' at mid-day : Sayana adds in explanation (not very 
clear in itself) that their milch-cows were kept at night in 
' salas,' but the rest of the cattle in the gostha.' Here 
' gostha ' cannot mean open pasturage ; and ' samgavim ' also 
seems to be ,sonie sort of an open shed where the noon-tide 
milking was done ; ' gostha ' and ' samgavim ' therefore would 
mean cowstalls and cattle-sheds attached to the clan-abodes 
and set up in the fields, respectively, while the * salas ' may 
have been special sheds for milch-cows with isolated compart- 
ments or each such cow may have been isolated in its separate 
' sal a.' 10 It would 'also appear that the * gostha ' belonged 
to the whole clan, e.g., of the Bharatas, 11 and not to the 

1 Geldner: Ved. ,Stud. 2, 275-276. 

2 Cf. Chand. Upan. IV, 4, 1 ; gankh. Sr. Sut. I, 4, 16; etc.; Asval. 
Grh. But. IV, 4; etc.; Kaus. Bra. XXV, 15; etc. (It_ is to be 

noted that Puranic tradition places the rise of noted c rsi gotras ' 
(clans) much earlier than the period indicated l>y these 
references. Thus the k clan ' sense is not a late one.) 

3 St. Pet. Diet., s.v. 

4 The suitability of such interpretation is evident in Rv. I, 51, 3; 

II, 17, 1^23, 18; III, 39, 4; 43 ; 7; VIII, 74, 5; X, 48, 2; 103, 7. 

5 Vide infra. 

6 Rv. I, 191, 4; VI, 23, 1; VIII. 43 3 17= Av. Ill, 14, 3.5.6; II. 26, 

2 = Vaja. Sam. Ill, 21 : V, 17. Kath. Sam. VIII. 7; Mait. Sara. 
IV, 2, 11 = Ait. Bra. HI. 18, 4; gat. Bra. XI. 8. 3, 2,; etc. 

7 Sp. in AY. Ill, 14. 

8 In Av. op. cit. 

9 Ait. Bra. Ill, 18, 14. 

in Of. ' s'alaV ante, p. 4, 11. 13. 

11 Examples of ruling and influential priestly families possessing large 
herds of cattle (often with special structures for these) S.T-Q well- 
known in Epic-Pnranic tradition, and the Bharatas are actually 
amongst them. 

( 8 ) 

individual houses or holdings; and it Is thus very probable that 
the craduallv more and more specialized social association and 
unit of the * gostlii/ often mentioned later, in Buddhist and 
classical literature, 1 grew out of the merry ckn-gatheniiRa at 
the ' gostha > in the evening after the day's toil aJid adventures 
in the fields and pastures. 2 

Just as the later ' club-house ' 3 (gosthi) was developed 
out of the common cattle-stand, so also some other types of 
associations and their suitable structures were closely connected 
with ordinary domestic conditions. ' Vidatha ' must be derived 
from 4 via/" to acquire, rather than from * vid,' to know, 
which gives the plausible meaning of something like the 
Witan to the * vidatha-/ but which can account for only a 
few of its many senses; ' vidatha ' therefore originally meant 
holding or house 4 ; but it is very often used in wider senses, 
involving the ideas of a larger structure and some sort ot 
assemblage. Thus in different passages Ludwig sees the 
sense of a sanctuary or- asylum, 5 and Zimmer that of iai smaller 
assembly than the * samiti ' 6 ; where a ' Samrat ' is spoken of 
as ' vidathya,' 7 the * vidatha ' must have been a royal estab- 
lishments a court or audience-hall ; where women are connected 
with the 'vidatha ' 8 (but not usually with the ' sabha '), it 
may mean & household assembly, social or religious, and the 
accommodation for such an assembly ; while Both makes out a 
reference to some secular wider type of assembly in many other 
passages, 9 ' Yidatha ' accordingly -stands for quite a variety of 
building structures, from probably the quadrangle or large hall 
of a homestead to specialized structures suitable for public use 
or court life. 

1 Various aspects of the e gosthi,' economic and social, are indicated 

and detailed in the early Pali texts, Kautilya, and Vats. Kfi. 
Sutra; the term has subsequently degenerated into the colloquial 
vernacular * gusth(t)L* 

2 The traditional picture of Krsna's early life (in some of the PuratjiaH 

also) is an illustration of how this development may luivc 
actually happened, 

5 This is the special sense in Buddhist and post-Maurvan literature 

(e.g. in Vats. Ka. Sut.). 
* J. Am. Or. S., 19, 12ff. ; c f. Rv X, 85, 26.27; I, 117, 25; II, 1, 6; 

Av, XVIII, 3, 70. 

5 Ludwig: Trans. Rv. 3, 261; cf. gat. Era. V, 3, 1, 13; Kat. fir. But, 

X\, 3, 35; also Ev. I, 31, 6; V, 62, 6; Ait. Bra- I, 30, 27.28. 

6 Zimmer: Alt. Leb. 177; Rv. H. 27, 12. 

! ? Y - ^V 27 4a 2; / L J'. 91 ' 20 > 167 ' 5 ' AV ' ^ 128. 
S Av. TO, 38, 4; Mait. Sai?lt ^ 7? 4; c 

AIV , 1, 20.25). 

9 BT ' x? 1 ! 1 ' 4 ; S- , m 7 : ? a ' 5 - 6; v - ' 

39, 1; X., 12, 7; Av. I, 13, 4; XVII, 1, IS. 

( 9 ) 

The well-known ' sabha ' is no less ambiguous in signifi- 
cance : the usually accepted view is that it denotes the ' as- 
sembly ' of the "Vedic Indians as well as the 4 hall ' where it 
met ; Hillebrandt however thinks that the ' sabha ' desig- 
nates primarily the ' house of as&embly ' ^iiile b samiti (also 
frequently occurring in Yedic literature) stands for the as- 
sembly Jl itself ; but it is noteworthy that while the " sabh* ' 
has a number of functions and aspects 2 ascribed to it in the 
Yedic literature, the most particular detail available about the 
* samiti ; is that kings and princes frequented them 3 : hence 
the ' samiti ' was more a political institution! than the ' sabha/ 
and of a select character, though the L vis ' are associated with 
both. 4 According to this view the 4 sabha ' would be the hall 
of the widest assembly of a. community and the ' vidatfaa ' the 

*-- i> 

quadrangle or hall of the unit family assembly ; and * samiti ' 
would have to be placed between these two types. In fact 
the ' vidatha ' does develop into the ' samiti ' type : for in 
some passages the ' vidatha ' may have the developed sense 
of a public sanctuary or asylum,' 5 and in some others 8 the 
4 samrat ' is ' vidathya ' or ' holding court/ in of course a 
suitable place : this latter use would correspond to the * .samiti ' 
associated chiefly with princes. In the Av. the ' sabha ' and 
the ' samiti ' are frequently mentioned together, 7 as equally 
ancient institutions 8 (where prepared speeches were made), 9 
which were to be found even in villages 10 ; while both were 
mainly composed of tribesmen and followed the King, the 
former was associated with the army, and the latter with 
' strong drink ' n ; and the ' sabha/ ' samiti/ and ' amantrana ' 
are mentioned as assembly -houses in order of increasing limita- 
tion. 12 On the whole therefore the ' samiti * seems to have been 
a. narrower institution. But there are other difficulties : there 
seem to have been several types of the s sabha * itself. Though 

1 Hillebrandt: Ved. Myth., 2, 124, note 6. 

2 Yide infra. 

3 Vide V.I., II, 4301. 

4 Av. IE, 19, 1; IX, 7, 9; XV, 9, 2.3. 

5 Vide ante. 

6 Vide ante. 

7 Vide following notes. Of. similar association in the modern 

vernacular phrase * sabha-samiti,' and its use in the sense of 
e wider assemblies and smaller committees.' ( So also the 
vernacular expression ' gosthl-gotra * affords a clue to fche 
relation between these twc parallel early institutions : ' gotra ' 
referring to the smaller unit of a family or * kin/ and * goRthi ' 
to the whole tribe or clan; in Buddhist and Maurya periods, 
the * gosthi ' is specially associated with the e gana,' which was 
wider than the 'gotra.'). 

8 Av. VII, 12, 1 (2 dtrs. of Prajapati, etc.). 
3 Av. XII, 1, 56; cf. VH, 12, 1. 

10 Av, Xn. 1, 56. 

11 Av. XV, 9, 2.3; (the context would show that the Av. regarded 

these institutions as originally derived from the e Vratya ' 
Kingship* of Magadha). 

12 Av. VIII, 10, 5.6 ; cf . Ev. I, 91, 20, whore a fit son is f sadanya,' 

'vidathya ' and t sabheva ' in increasing order of eminence. 

( 10 ) 

It is possible to conclude that all the multifold functions at- 
tributed to the ' sabha ' in different contexts were performed 
in one and the same institution and structure called ' sabha/ 
a state of affairs natural in primitive polity 1 (cf. Hellenic 
parallels!, yet it is reasonable to suppose that increasing com- 
plexity of functions very soon (even before the age of the later v 
Samhita^ led to a division into several correlated institutions 


also called ' .sabhas.' Thus, for example, the increase of gam- 
bling, so closely associated with the ' sabha ' from the very 
beginning, would in all likelihood lead to the growth of -a- type 
of special gambling halls, where this would not interfere with 
other more serious functions of the ' sabha ',; the everpresent 
and expert gamblers, the ' sabha-sthanus ' 2 would then leave 
the assessors, the ' sabhasads ' undisturbed in their judicial 
dignity; the two sets cannot very well be posited of the same 
hall at the same time. So also, we hear of the ' .sabhaviii,' 5 
the keeper of the gambling hall, a.s distinct from the ' sabha- 
pala,' 4 the warden of the assembly-hall; and of the ' graraya- 
vadin,' 5 the village judge or town-reeve, in his ' sabha/" or 
court, which is here apparently separate from the piniblm^ 
hall. Then again, certain other earlyWs of the word "' sabha'' 
would necessitate either a supposition that it was evolved out 
of domestic or individual household conditions, or one that we 
have in these instances a particular domestic use of the word, 
Thus when ' Agni ' of the ' .sabha ' is specially designated 
' vispah/ or master of the dwelling, 6 there is an evident re- 
ference to domestic conditions. In some passages in the later 
Samhitas (and subsequently) the sabha ' evidently refers 
to the ' society-room 5 in a private dwelling-house 7 ; and earlier 
still sabbeya' 8 and ' sabhavan rayih ' 9 seem to have been 
used domestically; while in ' sabhavatl yosa/ 10 of the Ev., 

!I - 56 - 7 V he same ' gathering (in the same hall) 
ooentk n FT? 8 and tbers v ' atchill S ' gae, is 
2hif 4 as - a OUrt f justice with lts ful1 



4 Taitt, Bra. Ill, 7! 4, 5 

; Mail, filup . n , 2j L 

4 ' 8 > 6 (a man's 

probably CMnd - Upan * 

S3 BI %J I I V/i ! jt T f\^ n*\ -- j - 

O AWV. ii, ji4, j_j- JL y]^ 2Q [nT-f)|-jQk]Q\ , A ,, YY too TT- <N 

XXH, 22. ipiuDaDiej, AV, XA, 128. 1; Vaja. Sam. 

q Rv IV '2 5 

' B ^^ISv^SlS" 

( 11 ) 

it is equally permissible to see in it a reference to the 
presence of vromen in the greater assemblies, the use is pro- 
bably a domestic one, meaning something like 4 the lady in the 
drawing-room.' It vrould tliu.-s appear that, whiclieve/ be the 
earlier model, the .sitting-room of a private home had much in 
common with the wider assembly hail, and that the ytiucture 
and equipment of the 4 ,sabha/ domestic or public, v;as of one 
and the same type originally. So also, both the central hall 
of a dwelling-house and the assembly hall had their fire-altars, 1 
the prototypes of the later ' worship-room ' ( L thakur-ghar ') in 
private houses, of the nave (' caitya ') in the Buddhist congre- 
gation halls, and of the sacred antechamber (' tkakur-dalln ') 
in assembly halls of all descriptions (e.g. the ' iiatya-sala ' or 
' nat-niandir ') : the difference being probably only 'in the size 
and type of the altar or other sacred" symbol and in the number 
and variety ot the k bthanus ' or pillars. The ' sabha ' in 
its wider sense must have been a large edifice with some pre- 
tensions to architecture ; apart from the altar 2 and pillars, 3 there 
must have been more or less suitable structural arrangements 
for the transaction of judicial, commercial and political busi- 
ness, and reception of courtly, well-born, wealthy persons and 
kings; and the complexity of the structure must have been 
greater where the same building was used for the other 
' ,sabha '-ic functions, gambling, merriment, social inter- 
course, debates and contests. 4 Probably when the social and 
festive branch of the ' sabha ' became separated it merged with 
the natural clan-gathering* at the k gostha,' and led to the 
formation of the later ' gosthi,' whose functions were pre- 
eminently social and pleasurable. 

Associations of learned men called ' parieads ' were in 
existence in the later Vedic period, 5 and the origin of this insti- 
tution may well be referred to the earlier epoch 6 ; at any rate 
these ' parisads ' were early converted into administrative insti- 
tutions (councils of judges and ministers \ and it is very pro- 
bable that the ' parisad ' either held its sittings in the 'tradi- 
tional ' sabha/ 7 or came to possess a. special habitation of its 

,- 4 '' V > 3 < U ' Vn < 7 > 5 >* ^- Vm, 10, 1-5; XIX, 55, 6. 
( Uns led to a metaphorical use of ' vi&'pati ' ; so also, apparently. 
the priest prayed at the ' subha ' altar while the King fought : 
V ! ? 11, 5) . 

2 Vide note 1 above. 

3 Vide ante, n 2, p. 10; so also in the Epic, pillars are the 

mam features of saLhas, while there are various adjuncts accord- 
ing to special needs and circumstances. 

4 For references for these several functions of the 'sabha' vide 

V.I., II, 426-427. 

5 Teni p Brhad. Upan. Jaim. Upan. Bra. and Gobli. Gr. Sut. ; vid 

Details m VI., I, 497. * 

6 This institution also was apparently originally of a pastoral 

character; the 'parisad/ rich in kine, is said to have been 

father s (Angirasas, etc.) for men: 
VJUI, 3, 22; cf. Ev, IV. 2 } 17. 

Which had its judicial side, 

( 12 ) 

own, As the ' parisads ' were mainly sittings around of 
' acaryas/ specialists in law and custom, sacred and secular, 
and as these- acaryas 3 had their ' kulas ' which were com- 
modioiis enough for resident students and their own families it 
would be quite natural for the sessions to have been held in 
some block of these ' kulas ' ordinarily. ' These ' acarya- 
kulas ' were not 1 merely one or two wretched huts 


(like their declining and impoverished modern repre- 
sentatives, the 'tols'). It seems probable that youths 
of all the classes of society were required 2 to, and 
even girls optionally could, 3 reside for a certain period in 
' brahinacarva,' though the period of such discipline may well 
have varied from class to class, and much of the course been 
optional or unnecessary for the non-brahman and girl pupils. 4 
These { kulas ' then must have been quite capacious and 
complex in plan. 5 A teacher might admit quite a 
number of pupils, 6 and Vedic as well as Epic-Puranic 
traditions refer to more or less specified numbers of 
resident students 7 in particular establishments. The 
* acaiya- ' was to teach everything to at least those staying 
on with him for a year, 8 while many students would stay on in 
their teacher's house for twelve to thirty-two years, even after 
the Yedas were done, 9 Hence the teacher pf the later Vedic 
period must have had in his ' kula ' sufficient accommodation 
of a permanent nature to provide for such prolonged stays and 
no doubt also frequent migrations 10 : such provision must have 
been possible largely through the voluntary fees of sons of 

3 Tims in Epic-Puranic tradition these are fully prosperous establish- 
ments, where princes are entertained sumptuously, and are 
quite comfortable places for them to be in residence for instruc- 

2 Re. probability of this system, vide V.I., II ; 75, 

3 As the application of ' brahmacarya ' to unmarried girls (who thus 

become fit for marriage) in the Av. shows, together with a 
number of actual cases known to Epic-Puranic tradition. 

4 Bnddmsb Bnraah still retains a trace, in its system of education, 

T7 ancient Indian theory and practice. 

5 Ilie Epic-Puranic accounts always depict them as such; cf. the 

description m Mbh. gaknntalop . 


reasonably numbers of residents said to have been 
TiJ v 6 - '-^j asr , aina ? b y tile ' Kalakeya raids of tradition. 
W b rs f , an earlier P eriod had full 

renowned professors* is^SOcP HUm ^ f studenls resid ent with 

8 SjjJb. JBtBi ^.X.j^'y 7 ^ 1 Ofi O? -,C A V T T-T . ~ 

gfii^ 1 ^ ^'l?- U: "^t'Upa 3 , I, B , 1 5 etc. 

'.S^iSS rrtVnt/^eciar,? in mig t h a e ti0nS ^ 
^"i-"*" - 1 -' Canted solutions of 

( 13 ). 

nobles and princes, 1 about which the Epic and other ancient 
traditions say a good deal. What the general plan 
of these * kulas ' were, we may gather from the terms 
' ante-vasin ' 2 and l acarya-kula-va.sin ' 3 used of the resident 
student : he dwelt near by, but in the outskirts as 
it were, yet it was all within the teacher's family 
home or establishment ; i.e. , the pupils' quarters were in separate 
blocks a little apart, which were still part of the same struc- 
tural unit. We might discover in these * acarya-kuias ' of the 
earlier epoch (residence in ' brahmacarya ' being known 
as early as the Atharva-veda 4 ) the same general plan which 
characterizes the later monastic establishments, 5 Buddhist or 
otherwise, a quadrangular structure with cells on all sides and 
the shrine and abbot's cell in the centre of the quad, or with 
the cells on three sides and the East-facing block set apart for 
the abbot and the shrine. The ' parisads * of learned men, 
therefore (and the similar but mainly theological associations 
of the ' upa-nisads/ or sacred and ' secret ' sessions to discuss 
the mysteries of theology) , together with the * acarya-kulas J 
(of which they were probably special developments) , may be 
looked upon as the later (or even early) Yedic beginnings, out 
of which the pre-Buddhistic and Buddhistic centres, of learning 
of the ' residential university ' type 6 were evolved. 

So also we find the prototypes of the Buddhistic trunk-roads 
and travellers* rest houses in Vedic conditions. Road-making 
indeed proceeded side by side 7 with the Aryan settlement : with 
reference to the extension of settlements in the new land, the 
clearing of forests, and making of roads, gods like Agni and 
Pusan, and ' rsis s (like the Roman ' pontifices *) are called 
' pathi-krt,' the path-makers. 5 The Vedic builders were not 
long content with forest-tracks or village -pa ths ; for even in 
the Rgveda (and later Samhitas) we find the ' prapatha ' or 
long journey by (broad) road, 9 and the Atharva-veda refers to 

1 Cf. the teacher's prayer in Taitt. Upan. I, 4, for material prosperity 

along with influx of large numbers of students. 

2 Vide n. 9, p. 12; also, Sat Bra- V, 1, 5, 17; Brhad. Upan. VI, 
A 3 } 7; Taitt. Upan. I, 3, 3; cf. Ait. Aran. ITT, 2 } 6; Sankh. 

Aran. VIII, 1, 1. 

3 Chand. Upan. II/ 23, 2 (settling long therein). Cf. note 4 below. 

4 ' Brahmacaryena vas ' : Av. VII, 109, 7 ; Ait. Bra. V, 14 ; cf . Av. 

XI, 5 (re the ' student'). 

5 E.g. as represented in the sculptures of the 3rd and 2nd centuries B ; C. 

6 Traces of whose elaborate structural arrangements are now being 

unearthed at the sites of Taksa-sila and Nalanda, 

7 If indeed roads were "not there already; the cross-country roads 

feeding the ancient S.W. seaports, may have been much older 
than Aryan settlement. [The Smdh-Punjab excavations of 1924 
seem to prove existence of such ports in the pre-Aryan India 
of the 3rd millennium B.C.] 

8 Vide refs. in VI., I, 489490. 

9 Ev. X, 17, 4.6; 63 ? 16; (cf. Ait. Bra. VII, 15). Kath. Sam, 

xxxvn, 14. 

( 14 ) 

the parirathva ' or road suitable for chariots. 1 ' Selu ' is 
found from the 5g- and 1 Yajur-vedas onwards, 2 but the precise 
s*n*e does not come out clearly. It has been held that a cause- 
war of an ordinary type, merely a raised bank lor crossing 
inundated land is meant, and that its ure is probably 
metaphorical in Teuic literature ; but a metaphorical use ot 
a term can hardly come into 'existence unless there has been 
previous simple "use of it, and the .sort of structure in- 
dicated here would be quite natural to and characteristic 
oi 4 the Gangetic delta, but can hardly be referred to 
the pre-eminently Yedic regions (or Madhyade&i) : besides, 
there is no inconsistency in ascribing to the ' setu ' the sense 
of a causeway ofi some '.special ' structure, a dam or a bridge 
(more of use in the Yedic regions proper) , when we find long 
road-journeys performed and drives constructed. 3 Later on 
(in the Brahmaiias) villages are connected with ' niaha- 
pathas ' or high" roads 4 ; and causeways C badvan ') firmer 
than an ordinary road are known. 5 A much earlier reference, 
to well-made pleasant cart-roads, on a higher level than adjoin- 
ing fields, forests and other village-tracks, with great Uw.w 
planted beside, passing through villages or towns, and with 
occasional pairs of pillars (i.e.,, gateways, evidently near the 
approaches of some town), is made in the Av., 6 where bridal 
processions pass through such routes. ' Prapalha, ' in the 
Yajur-veck has also the sense of a 'broad 1 road ' 7 ; while in 
Kv. itself ' piapathas ' are also rest-houses, apparently on the 
4 prapatha' or high road, for the travellers, where k khadi ' or 
food may be obtained 8 ; so also in the Av., where every ' tirtha ' 
along the bridal route is said to be well provided with drink, the 

1 Av. VIET, 8, 22. Whitney translates * rim. ' ; but * road ' givc k s a, 

better sense from the context, where a sacrifice is likened to a 
chariot journey. Cf. Av. XII, 1, 47 (many roads, for people 
to go upon, ( vartmans ' for chariots, and for the going of the 
cart, by which men good or bad go about, free from enemies and 
robbers; v. 45 refers to many countries with people of different 
speech and customs). 

2 Rv. IX, 41, 2; Taitt. Sam. Ill, 2, 2 3 1 : VI, 1, 4, 9; 5, 3, 3; VI 1, 

5, 8, 5; Kath. Sam. XXVII, 4; Ait, Bra. I [I, 35; Taitt V> ra- 
il, 4, 2, 6 : gat. Bra. XIII.' 2, 10. 1; Brhad. Upon. IV," 4, 4; 
Chand. Upan. VIII, 4, 1, 2; etc. 
Z Vide n. 8. p. 13, and n. 1 above. 

4 Ait. Bra. IV, 17. 8; Chand. Upan. VIII, 6, 2; (this agrees fully 

with early Buddhist references to such roads; vide "also ri. 6 

5 Pane. Bra. I. 1. 4; cf. Lat. Sr. Sut. I, 1, 23. 

6 Av. XIV. 1, 63 and XIV. 2, 6. 7. 8. 9. 12. Such a road is 

'ascended' from the village roads; it is possible thai, the 
f pillar standing in the way ' may refer to barrier posts, for 
the levying of toll or octroi on the trade routes. 

7 Kath. Sam. XXXVII, 14. 

8 Rv. l f 166. 9 (Wilson; Trans. Rv. 2, 151). The reading ' prapadeau ' 

is not necessary, as the connection between ' prapatha ' the high 
road and * prapatha ' the rest-house is quite clear, 

( 15 ) 

' tirthas ' are something like these ' prapatlias * being rest- 
houses on the fords. 1 The Av., and some Brahma-nas and 
Sutras, mention the L avasatha/ which, though literally mean- 
ing dwelling, is not used, in the general sense of abode till 
much later, 2 but which is used there in a- special sense, 3 a 
structure of some sort for the reception of guests, specially of 
brahman rf and others on the occasion of feasts and sacrifices ; 
it may have been something like the later k dharma-salas ' or 
guest- and rest-houses, though not necessarily on the high 
road. Travelling indeed seems to have been quite common : 
dwelling abroad and residence in foreign countries is mention- 
ed in the Ev. itself, 4 and the Av. has got its ceremonies for 
return from 4 pravasa ' (along with the Grhya Sutras), 5 and 
vividly describes the weary merchant's homecoming; while the 
Yv. Samhitas know of ' yayavaras ' 6 or travelling mendicants, 
probably the predecessors of the itinerant monks of the- 7th and 
6th centuries B.C.. The appellation ' Prapathin ' given to a 
'Ya'dava prince in the Ev. 7 may probably indicate that princes 
of those times, like their successors a few centuries later, were 
already makers of long roads and philanthropic rest-houses. 

Building-activities, indeed, developed in Vedic times not 
only through the needs of social and corporate life, as in the 
case of the ' gosthi,' the k vidatha," the * sabha, 1 and the like, 
but also fchrotfgh the kings and lesser chieftains. 8 In speaking 
of ancient Indian polity it is still customary to call up a vision 
of a sole monarch towering above a dead level of agricultural 
population ; but evidence for the Vedic and Buddhistic periods 
does not point 'to such Chaldaean simplicity. It rather appears 

1 Av. XIV, 2, 6. 

2 E.g. Ait. Upan. Ill, 12. 

3 Av. IX, 6, 5 (entertaining brahmans) ; Taitt, Bra. I, 1, 10,, 6; HI, 

7, 4, 6; gat. Bra. XII, 4, 4, 6;"Chand. Upan. IV, 1, 1; details 
in the Sutras : Apast. Sr. Sut. V. 9, 3; .Apast. Dli. Sut. II. 
9, 25, 4. 

4 Rv- VIII, 29, 8. 

5 Av. VII, 60. 1-6; cf. Asval. Grh. Sut, 1, 15; Sankh./Grh, Sut, 

II 17 etc. 

6 Taitt/ Sam. V," 2, 1, 7; Kath. Sam. XIX, 12. (The Epic tradition 

also assigns ' yayavara ' sects, to which Jarat-Karu belonged, to 
the period immediately after the close of the Bgvedic). 

7 Rv. VIT<1, 1 5 30 (the prince lauded for his superior weapons, horses 

and ' prapa.thas '). R is noteworthy that the name is given to 
a ' Yadava * prince, Asaiiga, who may bo placed at the close of 
the Rgvedic -period (being apparently the same as Asaiiga, the 
so'ii or grand son of Satrajit and a near relative of Krsna) ; 
tradition ascribes (cf. Mbh., Hariv. & Br.) much building 
activity in S.W. India to the Yadavas of the Rgvedic period, 
and all that is known of ancient commercial activities, points to 
the early development of communications in those regions. 

8 It would be most unusual, if they were not so developed. (Even the 

petty PUficala and other princes lauded in the Rv. were 
evidently opulent, and there were gre?ter and more famous kings 
than these). 

( 16 ) 

that between the King and the common people there were 
intermediate ranks of a. fighting nobility, analogous to the 
medieval knighthood of Europe or Rajput India J We must 
assume, for the Yedic (even Buddhistic) period, ^some ^such 
significance attaching to the well-known terms ' rajanya ' and 
'ksatriya' (and other cognate words). Apart from ^ this, it 
would appear that such a class is referred to in the ' ibhyas/ 
rich lords (in fact * ibhya ' later on 1 becomes a, synonym for 
rich and noble) , possessing retainers or elephants (privileges 
traditionally indicating lordliness) , 2 whom the- King is said 
to devour as fire the forest. 5 That the ' ibhyas ' were nobles is 
quite clear, but what ' ibba ' means is not equally so : Pischel 
and Geldner follow Sayana and Mahidhara's comments on the 
word in some passages in making it equivalent to elephants 4 ; 
but though this meaning is common later on, it is not so as we 
go back ; for the Nirukta gives both elephant and retainer as 
equally good meanings, 5 while the Asokan inscriptions have it 
in the -sense of ' vaisya ' or subordinate. 6 This latter use is 
significant for it shows that ' ibha ' really had a special political 
or constitutional meaning. Hence, in the Rgvedic and Yajur- 
vedic passages where it occurs, 7 it is better to take it in the 
sense of retainers and vassals, with Roth, Ludwig and 
Zininier 8 ; this entourage 9 may well have included, besides 
servants and dependents, members of the * ibhya'.s ' own 
family, and young cadefe from subordinate families of chief- 
tains (specially in the case of princes). 10 The existence of 
such lords is indicated also by the use of ' vefiajja^all.ih- 

1 Chand Upan. I 10, 1, 2; etc. (Vide V r p 80 , for other 
refs.). Even here the sense may be^ ' grl ma ' > belonging to an 
^ abhya or nobleman, and hence ' having K tainers an d elephants/ 

i.e., * rich, 

The Greek writers noted this for India, of theu t; mft . O t th<* 
story of King Dhrtarastra of the Knm rt m trating wi 

bmnmana for possessing an elephant. e 


5 Bv. I, 65, 4 (This relationship is a commonplace^ ^ < Eaiadharma ' 
tradition). . J 

4 Say, and MaM. on Taitt. Sam. I, 2, 14, 1 and VsL fs TO XTTT q- 

with Pischel and Geldner; Ved. Stud. I, xv-it- * ' > 

5 Nir. VI, 12. Y 1 ' 

5 Cf. Bflhler: Z.D.M.G., 37, 279, on Edict. No. 5 

? RT. 'I, 84, 17; IV, 4 1; IX, 57, 3 ; VI, 20, 8 (^ VeH 
name or title ' Smad-ibha * or Great Bran), ^ T^ 
1,2,14,1; Vija. Sam.. XIII, 9. * \ iait 

8 EotB: Diet., a.v.; Ludwl^ : Trans. Ev. 3 S 246-7; 2"; iirane r : Alt. 

tl>. Io7. 

9 Cf . the * apasti ' (comp- to the epic * .npastha * and\ medieval 

5 kayastlia f ) or dependents, clients proper of tfc e ^ri BS n ot 
servile, but specially related, as opposed to 'Ordinary L-ubiects 
including conquered tribal chiefs, ambitions men (like Sut 'as and 
Gramanis) and state officials. For references, vide V.I. 1 

6 Cf. voiiEg: princes of petty states in the entoiifage o! the b% 

King Javadratha. who serve him as standard-bearers, 

* w*$ % * 

etc., in Mbb, 

( 17 ) 

Sanihitas*- In the sense of vassal tenant or dependent neighbour : 
Geldner 2 Is content with the meaning of a neighbour or 
member of the same village community ; but this view is not 
tenable, as * vesya ' In Ev. 3 is used definitely in the sense of 
dependence, and * vaisya ' in Taitt. Sam. plainly means 
servitude 4 (besides other derivatives 5 used in the Samhitas with 
similar significance) ; again, the sense of neighbour belongs 
not to ' vesa ' by itse'lf but to ' prati-vesa/ 6 (also used in the 
Samhitas), literally 4 fellow- vassal/ hence a neighbour, the 
earlier word for it being * nahus/ of Indo-germanie origin. 
That vassalhood to a lord was not uncommon is Indicated by 
expressions (in the Atharva-veda and some Brahmanas) like 
1 natha-kama J or * natha-vid, * 7 referring to men seeking the 
protection of lords, probably much as the protection of Anglo- 
Saxon earls and Norman barons was sought by the ordinary 
freeholder or cultivator. 8 

Now it follows from all this, that from the early Vedic 
times onwards there existed something like a feudal military 
baronage, connected with kings on the one hand and dependent 
vassals on the other, wealthy enough to excite the cupidity of 
the former and enjoy princely prerogatives, and powerful 
enough to protect the vassals who sought them. All this how- 
ever would be impossible without something like baronial 
strongholds or other similar specialized structure. Evidently 
these are to be found in some at least of the Vedic * purs. 9 
According to this view the invocation of the king (in the 
Rajasuya) as ' puraip bhetta ' 9 gains appreciably in significance : 
an anti-baronial king fighting for suzerainty and order would 
certainly be better fitted for such eulogy than a simple l breaker 
or Backer of cities/ which would be more to Assyrian taste. 
So also this view gives a better meaning to ' pur-pati ' (of 
the Ev.) 10 than that of a ' a regular official, like lt gramani/* 

I Ev. XV, 3, 13; V, 85, 7; X, 49, 5 (prob.); V&j. Sam. (Kan) j H, 5, 
7 1 Mail. Sam. I, 4, 8; U, 3, 7; IV, 1, 13; AV. 33, 32, 5; vide 
also notes 3 6 below. 

* Gelduer: Ved. Stud. 3, 135 7 note 4. 

3 Ev. IV, 26, 3; VI, 61, 14. 

4 Taitt. Sam. n, 3, 7, I. 

5 ' Vesas * and ' parivesas ' in the sense of chief and subordinate 

tenants of the King, as opposed to * Ksdllakas ' or petty pro- 
prietors, in AY. II, 32, 5 j * vesatva * in Kath. Sam. XII, 5. 
(Of. St. Pet. Diet., s.v. 'vesa' and ( v&atva''). 

6 Ev. X, 66, 13: Taitt. Sam. II, 6 T 97: Vaj. Sam. XI, 75 j Katli- 

Sam, XXXVI, 9; Sat. 'Bra. IV, 1, 5, 2; Taitt. Upan. I, 4, 3. 

7 Av. XIII, 2, 37; XI, 1, 15; (cf. Pane. Bra. XIV, 11, 23); Av. IV, 

20, 9? IX, 2, 17; XVIII, 1. 13; Taitt. Bra. I, 6, 4, I. 

8 Cf. the Epic case of a robbed cattle-owner approaching Arjnna for 

protection ; and the epic maxim that first a * rajan ' is to be 
selected or chosen, then a home may be established, whr 
1 rajan * is rather such a baron than the * great king/ 

9 Vide VI.. II, 219, for refs. to * -raja-siiya ' passages. 

10 Ev. L 173, 10; (cf. Ludwig: Trans. Kv, 3, 204? and VX, 'H f 
1344.) ' " - 

( 18 ) 

in charge of a permanently fortified settlement/ or ' a 
temporary commander of a temporary fort or garrison ' (which 
latter is held to be more probable), 1 viz., ' lord of a castle/ 
an 'ibhya ' or 4 natha./ Such a view is further supported by 
the fact that spme jpf the ' purs ' had names ascribed to thern s 
such asPatharu, 2 Urjayanti, 3 or Narmini, 4 while some of these 
names were derived from those of chieftains possessing them, 
e.g. from Narmin(a} ? 5 or Sambara (his forts being called 
' Sambaras * in neuter plural) , 6 

* Purs ' were owned as often by the_ chiefs of the earlier 
population as by the new-coming Vedic Aryans ; Pipru of the 

" black brood ' possessed many forts, 7 and we hear of . the 
castles of Cunmri, Dhuni and "others, in all probability Dasa 
chiefs 8 ; while to Sambara the Dasa hero are ascribed 90, 99 
or 100 k purs/ 9 The real existence of the Dasa.s as 'a distinct 
people- in the Rgvedic times seems to be beyond doubt. The 
Dasas have their ' vis'ah/ and are classed as a ' varna ' n ; they 
were often dwellers in the mountainous regions 12 ; they had great 
wealth themselves, 13 and wealthy Aryan chiefs were" those \vho 
had * dasa-pravarga rayih ' or wealth consisting of troops of 

dasa slaves 14 ; and the women of the Dasas are "found as 
slave-girls and concubines. 15 It is thus quite unnecessary to 
take lipru Sambara -and others as other than real aboriginal 
but civilized Dasa chieftains, whom the Vedic immigrants had 
lound it not easy to dislodge from their numerous strongholds 
in the country. As however they were being ousted step by 
step, their forts would naturally pass into Aryan hands and 
become Aryan baronial strongholds, whence the * natha.8 ' ' and 

pur-pahs ' might protect the ' ve&s.' Sometimes ' purs ' 
may have formed parts of the ' gramas ' themselves 16 , ; in these 

1 The rarity of the word does not necessarily prove the tem.Dorarv 

/*O SiT*'Sf*TfiT' f^T TiliO fltTnm Trv* n * *3 * "* _L * Tl 'Jt **)* 

cDaiacier or tne command; it is equally accounted for bv the 

na.T.nT*allTr fh^ rii -wnnl/l k. !* -T .TT 

ibis wouia oe i-ess lanciihar "with thp 
' gramaiii)^ 

takes it 

Bv> 3> 152) * 

148'. SBE a i) St '. Pet Dick "- (OWenberg: Kv.-noten, I, 

6 V.B., II, 355. ? 

7 Rv. I, 51. 5 : VI, 20. 7. 

8 R %TV 8 i5 8 K-JS; $ $> 26 > 6; IV - "> , .- x, us.. 

''!' ' ?; H ' }9 ' 6; U > 14 ' 6 5 "> 24, 2. 

10 Jfe S*mbar as a real Dasa, of. Rv. 1. 130, 7 IV 30 14 VI Pfi S 

U For referees vide V.I.. I, 356-358. ' ' ' ' YI> ' 26 ' 5 " 

2 I ' * - - vn, . , 

Ait - 

< 19 ) 


cases a whole clan or band of Arvans instead of mightv chiefs 

**' C" ts 

may nave overpowered and entered into possession of some 
minor Dasa stronghold, and then made it the basis of their 
1 grama ' settlement. On the whole the view of Zimmer, and 
others after him, 1 that Vedic India knew of nothing more 
solid and -complex than the hamlet, like the early Germans 
and Slavs who had no castle-structures and town-life, is an 
extreme one ; for it is now being realized more and more as a 
basic fact that the- Vedic Indians, like the Iranians, Hellenes 
and Italians, were .superimposed upon an earlier civilization, 2 
in all probability of the same type (and maybe of cognate 
origins) as in the other three cases, and were similarly affected 
as regards religion, arts and crafts. 3 The Germanic" parallels 1 - 
therefore should not be carried too far. 4 Thus it becomes quite 
reasonable to find in ' prthvi,' 5 l urvi,' 5 ' ' satabhuji/ 6 
' a&namayi,' 7 or ' ayasi ' 8 ' purs/ or the massive, extensive. 
hundted- walled, stone-built, or iron -protected .forts, vivid 
descriptions of new and wonderful things the Vedic heroes 
actually saw ; and the rather forced explanations discovering in 
them mysteries of myths and fancies of metaphor become un- 
necessary^ The main difference, originally, between the Dasa 
and the Aryan * purs * must have lain in the materials used 
(which depended on the nature of the country they were 
familiar with) , large sections of the former being acquainted 
with the Vindhyan and Central Indian granites 9 and metal 
ore?, 10 the latter with timber -work mainly. But adaptations 
from one another seem to have occurred quite early: Susna. 
apparently a Dasa> enemy, used ' put carisnu/ or small 

1 Summarized in V.I., I, 538540. 

2 [Extensive remains of this earlier Indian civilisation (cir. 3,000 to 

2,000 B.C.) have very recently been discovered in the Punjab 
and Sindh. Many of the suggestions and inferences in this workj 
based upon literary evidence chiefly, will be found to be remark' 
ably corroborated by these archaeological discoveries. These also 
make it -almost certain that the W. Asiatic or Minoan civilisations 
had much in common with this earliest Indian civilisation which 
was thefr source both racially and culturally. This field of inves* 
tigation promises to be most fruitful for PurSrdc scholars and 
epigr aphis ts.] 

3 Thus it is demonstrable from traditional accounts that Vedic Brah- 

manism itself was originally non- Aryan (cf. Pargiter: AIHT)* 

4 There is really very little of common conditions. 

5 Rv. I, 189, 2. 

6 Rv. I, 166, 8: VII, 15, 14. 

7 Rv. IV, 30, 20. 

8 R V . I, 58, 8; II, 10, 18; 20, 8; IV, 27, 1; VII, 3, 7; 15, 4; 95, 1; 

X, 101, 8 (cf. Muir. Sans. T. 22, 578ff.). 

9 Cf. n. 12, p. 18,- (the hill-tracts referred to would appear to be 

mainly Vindhyan, if the traditions regarding the distribution of 
pre-Aila races are taken along with it; so also according to these 
traditions the Alias came through North Himalayan regions into 
the plains just below, an area still famous for timber art and 

10 Iron and copper smelting by using surface-coal is almost a pre- 

historic achievement of 'the Dravido-Kolarian races of if-E. 

( 20 ) 

moveable forts/ evidently constructed of timber ; it could only 
have been either erected on trucks with four or more wheels 2 
to be drawn by horses or elephants, or composed of adjusted 
parts easily dismantled or pit together, a ort of 
* camp-tower ' 3 ; so also the Aryans had their ' pa?ya 
or stone-bulwarks, 4 but the use of this word to denote 
also the stone slabs for pressing / soma/ 5 ^ shows that 
such defences were a later acquisition. * Dehi/ a defensive 
construction of some sort, 6 is used specially of non- 
Aryan defences, though not invariably; it might mean 
either hasty defences thrown up against an enemy, or 
more permanent earthworks and dykes, or -rubble rampart and 
trench going together, which last Is the most suitable sense. 
It is likely that these ' dehis * are the '"'saradl ' 7 or * autumnal ' 
forts ascribed apparently to the Dasas 8 ; these may have /been 
more or less temporary earthworks^ ramparts or trenches, 
constructed every autumn to meet fresh , campaigns of the 
Aryans 9 ; but in the course of time ' 'dehis * found to be of 
strategical service would come to be permanently used. 19 

1 Rv. VIII, 1, 28 (cf . Hillebrandt ; Ved. Myth. 1, 300 n. ; 3, 289 n.<). 

2 Like the later ' rathas/ e.g. as represented at ^Kon&rak. Cf. 

teniporary residences, like * rathas,' built for kings in Bajasuya. 
sacrificial area, in Mbh. 

3 The construction may have beea suggested by the * ratha-vahanaB ' in 

common use in the Vedic age, which were moveable stands for 
chariots, probably drawn by horses ( * ratha-vahana-vahas " ) into 
the battle-field, where the chariots were then used in. action, 
With this may be compared the many-wheeled stands used in 
the same way even in the present day for the * divine ' * rathas * j 
these ' rathas ' on stands indicate what th * pur carignu ' must 
have been like. This makes it probable- that the references to 
more than 4 wheels for a chariot are not mythical in every ?ase. 
Thus something like a many-wheeled ' pur carin'U ' QeGma to 

have been thought of in Av. X, 2, 28ft, where the * pur * of 
* Brahman ' is described as 8-wheeled and 9-doord. For 
1 ratha-vahana,' vide : Bv. VI, 75, 8; Av. Ill, 17, 3; Taitt. 
Sam. IV, 2, 5, 5 -Kath. Sam. XVI, llMait. Sam. II, 7, 
12^Vas. Dh. Sut, H, 34, 35.' Of. also, Kath. Sam. iXI, 10 j 
Taitt. Bra, I, 7, 9, 6 : gat Bra, V, 4, 3, * 23ff. " For * ratha- 
vahana-vaha/ vide t Taitt. Sam. I, ,8, 20, 1; Taitt. Bra. 
T, 8, 4,3; Kath. Sam. XV, 9; Mait Sam. II, 2, 1. 
* Bv. I, 56j 6. 

5 Kv. TX, 102. 2 MacdoneU : J.B.A.S. 1893, 457-^58). 

6 Rv. VI, 47, 2; VII, 6, 5 (cf. Schrader : Preh. Ant, 344; Zinimer : 

Alt. Leb. 343). 

7 Bv. I, 131, 4; 174, 2; VI, 20, 10. 

8 Bv. I, 103, 3; III, 12, 6; IV, 32, 10. 

It may be possible to connect c dehi ' with * dih,' to smear or 

plaster, and thus to take it as a mud wall; but it is noteworthy 

that *d(d)ihi' ' d(4}ih,* ' dah ' or Vda/ are quite common-place 

name_s in Bengal, Bihar and Chotanaerpur (-regions where indigenous 

non- Aryan elements are often clearly traceable), have a similar 

implication of trench and ramparts, or a defensible area of Jiigh 

nigged eronnd (cf, the-E. vern. xpr, * dah pa<Ja } * to get a wound 

like * ditch and wall '). Probably the ancient place-name 

s Vi-delia (gha) * is to be traced from a "Qehr fort: cf. - dala * 

in ^Vi-saV *Vai^ali > or f VWfila, 1 in the same region. 

M> Thus giving rise to place-names with * dehi * or its cognate words 

(and possibly even with ' pur *), 

Autumn indeed has always been the traditional season for 
military ventures in India, when the rains cease and the 
country becomes fit for marches, and the tradition 
probably goes back to pre-Aryan experience; it is 
difficult to see the point of the usual explanation 1 that these 
structures were intended to afford shelter from the ' autumnal 
inundations' and were therefore "of the nature of dykes. 2 On 
the other hand the ' purs ' which might, like fort Patliaru, 
be .saved by rain-storms from being set on fire, 3 or in the siege 
of which fire was used, 4 or again, whicE were full of 
kine f goniati s ) } 5 were evidently timber-built and characteris- 
tically, Aryan. 6 - TheJVedic 4 gomati purs ' -are the prototypes 
(or paratypes) of the Epic ' go-grhas/ or fortified, extensive, 
oowstalls, , the scenes of many knightly ventures, and 
possibly the ' go-pura,s ' of later architecture 7 are to be traced 
to this origin. The * gomati purs ' must .have originally been 
protected merely by earthen ramparts, with timber palisade and 
ditch. In some cases the' palisade of an Aryan ' pur ' may have 
been only a hedge of thorn or a row of stakes 8 fixed vertically 
and horizontally, 8 serving to make the approach difficult for 
enemies : the Egvedic ' durga ' 10 may -have primarily meant 
some such ' pur/ with thorn-hedge, stakes and ditches as 
hindrances to approach, but the meaning of a regular fort or 
stronghold may suit the passages equally well. 11 ' Vapra,' so 
frequent later on, occurs in the Av. in the sense of rampart, 12 

1 E.g. In V.I. 

2 For, firstly, no floods usually occur in the autumn; secondly, these 

floods are not formidable in Madhyadesa. 

3 Ev. I, 112, 7. 

4 Bv. VII, 5, 3. 

5 Av. VIU, 6, 23. 

6 Cf. the Epic (Bharatan) * go-grhas,' and the arrangements for the 

cattle of the Bharata clan (in Ait. Bra.; vide ante.). 

7 The association of the * divine bull * with later ' go-puras * mav not 

"^tt OJt v 

be accidental. 

8 Cf. Bv. X, 101, 8 (Zimmer: Alt. Leb. 143445); also Bv. VHL 

53, 5 (Both : 2. D. M. G., 48, 109). 

8 This earlier fencing is represented in the later * sala-protected * 
cities known to the Upanisads, and in the toassiv& Maurynn 
timber-palisades and stone railings, 

10 Rv. V, 34, 7; VII, 25, 2, 

It In Rv. X, 85, 32=Av. XIV, 1, 64, 'dtirga* (difficult of approach 
and reached or passed by * suga * ways) is used in a manner 
that^ indicates acquaintance with campaigns amidst hill-forts, 
Ancient place-names with * durga ' (' durg ' or * drug ') are 
found chiefly in Central India and S. W. Deccan, and these are 
of strong rock-fortresses; this might throw some light on the 
type of forts meant by the Vedic * durga.' Probably the 
epithets ' durgaha ' (unapproachable) and ' giriksit * (rock-render 
or rock-dweller) given to Mandhatr or other princes of his 
line) refer to such forts, sp. as ace. to Pur. tradition, Purukutsa 
and his brothers etc., are connected with Deccan expeditions, 
and Man4hatr also came into close contact with the S. W. 
Yadavas etc, (cf. also the Iksvaku kingdom in the Narmada 
region, and the place-name Mandhata '=anc. Mahismati). 

12 Av, VII, 71, 1 (Whitney: Trans, Av. 435-436). 

but the reading Is somewhat doubtful; while the equally 
familiar ' prakara' occurs only in the Sutras, 1 and IK used to 
denote a walled mound supporting either a platform <uud gallery 
for spectators, or a palace (' prasada ') 

' Pur ' and ' pura ' in Vedic literature arc probably ? not 
identical, as they are usually taken to be. ' Pura \ in 
' tripura ' 2 and ' mahapura, 3 occurring in the Yv. Saiplutaa 
and Brahmanas, is evidently something much bi^or : the 
reference is to cities with three * purs ' or three rows oC forti- 
fications and to great fortified cities, rather than to an ordinary 
* pur ' or fort with three concentric walk, and to a big fort 
only. 4 This form * pura/ again, occurs from the time of the 
Yvl onwards, when capital cities like Kamplla, had become 
familiar to brahmans; it is probable, however, that we have 
this form earlier 'still in the Ev., in the proper nanien 
1 Purarndhi ' and k Puraya/ 5 which, like the name ' Naga-rin ' 
in the Brahmanas, 6 may indirectly point to the existence of 
such ' puras ' or cities 7 in the earlier period. On the ground 

1 Sankh. Si-- Sut. XVI, 18, 14. (Tbese stages may however only 

indicate the gradually growing familiarity of bra-lunays as a 
whole with a pro-existent court and city life; which was 
clearly a late Yedic feature), 

2 Tait:. Sam. VI, 2, 3; Kath, Sam. XXIV, 10; etc. Sat. Bra. VI, 

3. 3, 25; Ait. Bra. II. 13 3 Kans. Bra. (in Ind. Stud, 2, 310). 

3 Taifct. Sam. VI, 2, 3. 1;' Kath. Sam. XXIV, 10; Mait. Sam. HI, 

8, lj Ait. Bra. I, 23, 2; Gop. Bra, H, 2, 7. 

4 'Trip-lira' is actually the name of a N. W. Deccan city in Pur. 

tradition ; so is * Satpura J in the same region ; both ' connected 
with much fighting and romantic tales regarding the Yadavas 
and their hostile neighbours, (cf. ' Dasa-pura/ also in the same 

5 BY. I, 116, 13; VI, 63, 9; ('puramdhi' occurs in other senses in 

AY. XIV, 50 ; Bv. I, 134, 3 ; Taitt. Sam. VII, 5, 18 , etc. ; vido 

fi Ait. Bra. V, 30; Jaim. Upan. Bra. HI, 40, 2. 

7 f PuramdM J is explained by Sayana, as 'of great dhi '(!),' and he 
takes * vadhrimati ' as a proper name (which is unlikely) ; as 
a princess is referred to, ' pura ' in ( puramdhi ' may appro- 
priately be taken to mean 'city'; so also" with * pura ' in 
* purava,' the name of a king (who gives away horses, slaves, 
cars, and ' pakva,' or brick-built houses). Proper names with 
pura, are not .uncommon in the Pur. dynastic lists. For the 
form ' puramdhi/ f. th$ later ( puramdhn.' ( (P.uramdhi ; seems 
to have meant "residing within a e pura ' or fortified capital,'* 
i.e. a noblewoman or princess, such as ' Vadhrimati ' was ; for 
, forfc of designation cf. ' Subhadra KSmpllavasini ' of Yv. 
and Snbjiadra Bvarakavasinr of the. Epic. Keith translates 
puramdh m Taitt. Sam, YET, 5, 18, by 'prolific woman 1 ; 
but as the prayer there is for 'this kingdom,' where the birth 

SSL, prjnce i an T a ? cll ? r 5 a hel 'o< a rathl* and a ' saldieya ' 
>outi2 is also desired, 'puramdhi' in this group must cor- 
respond to sabheya 3 and mean what was later called" 

Tl. in* irt** *T. a f _ f T^J *V ^ ^f it *f i 

ci. M.V. i, iq } 3^ where a * puramdhi * maiden 
at ?tf ^ ^ cr ^ over 's visit. In the Av, (XIV, 
- a / ^rarndhi J is. invoked in the ~ marriage rites, 
t prolific woman * might suit, but it Is more 

,f^ : ,* e ;!..- tb ^_i ua ^!?' a ***. de % "f >e 

the Epic) a 

23 y 

of the late occurrence of ' nagara * it has been held that city- 
life was not developed in Vedic period, -and that possibly ^there 
were no towns. 1 But ' nagara,' city, occurs definitely in an 
Aranyaka, 2 which means a good deal, as it implies that the 
fame* of the * nagara J was wide and longstanding enough to 
have awakened interest even among the brahnians in the 
s aranyas ' 3 ; then again, it is quite clear from the occurrence of 
Na-garin/ 4 resident of a ' nagara ' or capital city, as a proper 
name, and of ' Kausambeya ' (native of- Kausambl city) as an 
epithet, 5 that cities were in existence in the earlier Brahmana 
period. But at this point we lose sight of the * nagara.' At 
the same time, from the Brahmanas backwards up to the TV. 
Saipkitas, we find a substitute, the * pura/ 6 while we also 
get well-known names of cities for the period- 7 Going further 
back, the city is no longer to be distinguished as such, but 
still there is the * pur/ ' durga/ and other cognate settlements 
involving many 'different structural types and grades. The 
inevitable conclusion is that the c pur ' is the prototype, 8 the 

* pura ' is the developed city, and the * nagara 9 is the full- 
fledged capital city. It is to be noted that the sense of any 
ordinary town for c nagara ' is quite a modern one ; even in 
classical literature * nagisra * always stands for the imperial 
capital, at any rate one claiming such status or traditions. 
This makes it quite probable that the first occurrence of 

* nagara * in the Brahmana and Aranyaka age does not mean 
the first coming into existence of towns, but simply marks a 
stage in the history of Indian cities 9 and of the struggle for 
overlordship among the principalities and peoples of Northern 
India following Vedic settlement, the principal * pura ' of the 
paramount tribe or state being designated * nagara/ like 
k naga ' or rocks, 10 by way of pre-eminence in strength, 
or probably by way of reference to its stone walls or 
towers. Tlae references in the Upanisa-ds to 11- or 9-gated 

1 Vide V.I., I, 558-540. 

2 Taitfc. Aran. I, 11, 18; 31, 4. 

3 Cf. the brahmanieal notice of Ayodbya as a e grama.' 

4 Vide ante. 

5 gat. Bra. XU, 2, 2, 13; Gop, Bra. I, 2, 24. 

6 Vide ante. 

7 E.g. Kampila, Asandivantj Vara^avat! (Av. IV, 7, 1), or Kaus'ambi 


8 Also a \vider class; * purs * existed in the time of Bra.s and 

Upan.s also; e.g. Taitt. Bra. I, 7, 7, 5: Ait. Bra. I, 23; II 3 
11; Sat. Bra. Ill, 4, 4, 3; VI, 3, 3, 25; XI, 1, 1, 2-3; Chtod. 
Upan. VIII, 5, 3; etc, 

9 Compare the account given in Mark. Pur. (xlix, 41ff-) of the 

development of civilization. Here the ' pura * (big fortified 
town) is regarded as succeeding * fortresses ' in time, and 
preceding < the royal capital * nagara/ 

10 Capital cities and royal castles (e.g. descr. of Indrapraslh*) are 
always compared to rocks and peaks in the Epic. 

( 24 ) 

citadels 1 'thus reveals a new appropriateness, in the compari- 
son of the proud and striving ' bodies ' of the individual and 
of the corporate tribe ; it becomes unnecessary to see in such 
* poras * mere forts, and then to hold that 9" or 11 gateways 
are fanciful, their number depending on the nature of the body 
which is compared 2 ; no doubt only one gate in a city is men- 
tioned in a Brahmana, 3 but a comparison with 9- or 11 -gated 
cities could hardly have occurred to people who had never 1 seen 
more gate to a city ;, they may not have seen precisely 
9 or 11 gates, but any other number, say 8 or 12, which i's 
more probable, 4 as the earliest references to town plains, e.g, 
in the Manasara, Megasthenes, or actual remains, while they 
are all subsequent to the period in question ,T all point to the 
number of gateways being 4, 8, or multiples of 4 even up to 

The capital city, * pura ' or ' nagara ' must have belonged 
to some king or ruling family ; -and we should expect to find 
ample references to the special edifices connected with them - 
but such allusions are rather general and meagre, until we come 
to the close of the Vedic period. It is not that court and, city 
Me did not exist in the Vedic age, while it did in the Epic; 
is, rather a superficial appearance due to the - fact that the 
i 5? ^t 161 P? 65 * 1 ? literature had much less to do with court 
than the epics and the Ksatriya traditions .toad,; thus, when 
we come to special sections pf the Yv. Samhitas, 7 which 

forth S b S m n thi , ngs [ ega1 ' some more <ktwls do come 
forth. Ihe Bv. kn<ras : of such a thing as a King's palace and 

Varuna has one.* The ' harmya,' primarily b deSottoS he 
Vedio house as a unity, including stables, <Jc. , very soon 

1 Kath UpaB. V, 1 (11) ; gvetal Upan III 18 (9); (of. Schrader : 

. nS$.^'^iSft 451rWeber 

2 Kfttth: Ait. Iran. 185. 

5 at Bra, XI, 1/1, 2, 3. 



_ f " "-" J *if * W _B M * Ml jT 

fflr 'S fl TT J VrV^w \/4 

Jro Er^'TVbJr^^S frvl irif j * 

BV. 1^;^^?%^ f i 1 *?" "^ 

etc,-Rv. Vn > '56?l 6 4 ; 'cf 8> X %<' g 43 ' 3; 73 ' 10; etc - **t 

( 25 ) 

added on the qualification of being protected by a palisade,, or 
wall 1 ; and in the Ev. it-self we find a ' harmyesthah. " prince 
standing probably on the roof, or rather the balcony, of his 
palace, 2 just as any later Indian king would do to please his 
people. When the Av. thinks of a residence for Yama, it is 
a * barmy a.' 5 The specialized structure of ' prasada J is how* 
ever, explicitly referred to rather late in the post-Yedie 
literature. 4 But it is clearly, indicated in the earlier pccur- 
rence of * ekavesman/ 5 the towering prominent abode of the 
king as contrasted with the numerous houses of the people. 

A quadrangular style of palace-structure (comparable with 
the old town-plan of 4 roads and gateways or multiples of them) 
Is known as a main primary type in the Puranas (which appear 
'to have got, their technical information in common with the 
4 Silpa-sastras ' from some earlier special treatise, and whose 
compilers, the Butas, were also specialist builders to kings) , b 
and this is termed ' vairaja.' 7 It is perhaps pertinent to .see 
in ' vairaja 3 a reference to the .sort of k harmyas ' or residences 
toe early Vedic chiefs raised for themselves on, attainment- of 
4 vairajya ' or paramount cy of some sort ; l viraj ' is a royal 
title in the Rv. and Av., 8 and is well recognized in Pirranic 
tradition; but in Ait. Bra/, it is said at- that time to have been 
used in Uttara-Kuna and Uttara-Madra only 9 ; hence > either the 
4 vadraja ' type of palace-construction (known to the Puranas) 
was introduced into Indian Midlands (in the 8th century B.C.) 
from the,se Himalayan regions (whence the model form of 
Sanskrit speech also was derived in that age), or the style 

1 Rv. VII, 55, 6. 

2 Kv. VII, 56, 16 (Geldner : Ve"d. Stud. 2, 278, n, 2; Alt, Leb. 149). 

3 Av. XVm, 4, 55. 

4 Adbhuta Bra., in Ind. Stud. 1, 40; cf. prakarV and ' prasada " 

rising on it ; Sankh. Sr. Sut XVI, 18, 14. 

5 gat. Bra., I, 3, 2, 14. 

6 Thus the chief architect to Janamejaya HI (the Great) was a 

Pauranika Suta (Mbh.)- It probably indicates that palace 
architecture and fortifications were pre-eminently a Magadhan 

7 Gar. Pur, XLVII, 19ff. (re palaces). 
6 For refs. vide V.I., II, 304. 

Ait. Bra. VIII, 14, 3 5 this particular seems to be historically 

significant, as in the time of the Ait, Bra. (Tide Pargiter : AIHT, 

326, etc.) the (Southern) Kurus and Madras had ceased to exist 

as kingdoms, the former uniting -with the Pancalaa and ever 

retreating eastwards, the latter being lost altogether; the Kuni- 

Pancala, Kings used the simple title of ' raja/ as compared with 

the Eastern rulers, showing that they had decayed considerably- 

( 26 ) 

was a more ancient one, continued even after the passing away 
of * virat '-ships of the early Vedic and Epic period - 

Some details regarding the Vedic Kings' pal ace^ occur 
incidentally in connection with the court ceremonial of 
Rajasuya/ During this the ' ratna-havis ' rite wa,s performed 
at- the "houses of the King's ' Batnins,' something like a 
cabinet of King's Friends, including the chief Queen and the 
Household Officer *r These Ratnins' houses must have been 
round about or adjacent to the King's palace, 3 being in the 
same royal and sacrificial area ; and the separate houses of the 
sacrificing King's ' mahi?i,' * vavata/ and ' parivrkti,' indicate 
the existence of a complex palace of the harem type. It is 
noteworthy that both these particulars are borne out by the 
details of the Mahabharatan court, traditionally assigned to 
about the same period as the compilation of the Yv. 
Samhitas.* Of the different offices a ' Ksattr ' at the royal court 
might fill, 5 the Satapatha names that of ' anlahpuradhyak$a ' 
or ' harem-superintendent ' (which might be polished into 
* chamberlain '), thus implying a full-fledged palace establish- 
ment. This is also indicated by the other alternative (unctions 
of the Ksattr, who might be the ' gate-keeper ' (of the palace) , 
with assistants called ' annksattr,' 6 or the ' distributor^)! the 
King's gifts, etc. ' Here also the epic accounts agree afto the 
functions of the Ksattr, and the elaborate court-life implied. 7 
Thus it may reasonably be concluded that what is hinted at 
in the meagre references of the priestly texts is only given hi 
full in the Epic, quite naturally, It is also noteworthy thai 
chiefly those details re royal establishments are given in the 
former, with which a sacrifidial priesthood would be most 

itt. Sam. I 8, 9 Iff; Taitt, Bra. I, 7, 31f; Sat, Bra. Y, 3, 1, 
IffMart. Sam. II, 6, 5; IV, 3, 8j KSth. Sam, XV, 4. 


* f aS , in , 

pUlls and ytolisms xv ere adapted 

f Chief advisers ** >W were 


otT/ aaCe ai '" a j al n& willi tbe separate . 
of the chief queens and princesses, 

V ide Pargiter: AIHT, pp. 318, '20, '21, '23, '24 etc 

& a 

( 2? ) 

acquainted; thus, again, Janamejaya~Pariksita''s capital is 
called by the general epithet of ' Asaadivant ' (possessing the 
throne), 1 instead of the famous Hastinapura, showing that 
these brahraans were usually shown into a ' throne '-room or 
audience-hall of the King, and that was all that they saw of 
the court ; the other things striking them being, the awful 
' gate-keeper 3 with his staff, the royal disburse! of gifts which 
they appreciated, and the 4 harem-superintendent ' who con- 
veyed to them reverence and presents from the court ladies- 2 

One of the King's ' council \ of ' Eatnins ' was a 
Gramani ' which post was the highest ambition of the pros- 
perous ' Vaisyas 1S ; he may have been elected or nominated 
from the many ' gramanis ' of the state. 4 This makes it 
quite possible that through these selected and aspiring 
' gramanis ' imitations of the royal court, and its style and 
structures spread into their respective 4 gramas " and ' uiaha- 
gramaa ' 5 (the bigger villages or townships). 6 Thus a 
* grama ' also had its ' sabha/ 7 where the ' gramya-vadin f 
held court ; some l gramas ' may also have had ' purs/ 8 where 
the ' par-pati,' a ' natha- ' or w ibhya/ would play the king, 

Apart from these, the ' gramas J must have had other 
constructive activities (individual or joint), of maintaining 

1 Ait Bra, VIII, 21: gat. Bra. XIII, 5, 4, 2; cf. gankb. gr- Sui. 

"XVI, 9, 1. 

2 These points are copiously illustrated in all traditional stories 

regarding the connections between brahmans or ris and the 

3 Taitt. Sain. II, 5. 4, 4 5 Mait, Sam. I, 6, 5; cf. Weber : End. Stud, 

10, 20 3 n. 2; also Sat. Bra- V, 3, 1, 5. For references to the 
post of ' gramani ' in BV-, Yv., Av. and Bra.a, vide V.I., I, 
247, n, 25-28 and 31. 

4 Ifc is however possible that the ' gramani ' hera is ihe * mayor o! 

the capital city*; if Ayodhya could be called a * grama/ a 
city official also could be called a ' gramani ' by retired ri s 
knowing no better. 

5 Jaim. Upan, Bra. Ill, 13, 4. 

6 The idea of introducing styles of the capital city Into other towns 

and villages Is quite ancient, being referred to in the Tats Ka. 
Sutra as one of the primary functions of the metropolitan 
' gos(hls ' (a much earlier institution). 

7 Also ' samiti ' and ' amantrana ;' vide ante. 

S The ancient Katriya ballads (e.g. re Prthu) in the Pur- mention 
1 grSma-c&rgas ' as unnecessary or disappearing under a strong 
ideal king; these would thus seem to have been something 
* adulterine castles.' 

( 28 ) 

the interconnecting 1 roads, 2 or setting up grain-stores. 3 Defer- 
ences to structural forms in the Veclic village are in fact fairly 

' " Thus 


(granary) indicate gi 
for the earlier period also. The village well (' avata,' 8 
* kupa ' 9 ) had already its mechanism of \vater-w heel.B, 10 etc, ; 
and dams 11 ( 4 vartra s ) 12 were constructed to form tanks. 13 
These structures could not have been of ia rude primitive type, 
as the Aryans must have found these agricultural and irriga.- 
tional arrangements already fully developed in the Dravidian 
village communities, 14 

The house construction outlined in the Atharvaveclsi 15 
evidently refers to the ordinary type of dwelling-house in a 
village settlement, such as a brahman would 'either himself 
possess, 16 or consecrate -with mantras for the villagers under hia 
ministration. Such ji house was apparently characterized by 
these features: (lp ' Upamit' s ' pratimit's and ' parimitV: 
which seem to mean timber pillars and beams, in various 


1 Chand. Upan, VIII, 6, 2. 

2 In the Jatakas irrigation works are the joint concern of more than 

one settlement (e.g. of Sakyas and Koliyas), 

s Brhad. Upan. VI, 3, 13. These grain-stores must have been (as 
they are even now sometimes) quite large and complex structures, 
of timber and bamboo, plastered walla and raised platform or 
stone bases, Cylindrical, and, with round dome-shaped top ; a late 
medieval brick and atone model of such a capacious * gola * 
(* round 5 ) is the famous imposing ' gol-ghar ' of Pa^na; there 
may have been other masonry * golas r in earlier times also ; it is 
very likely that one of the sources of the * stupa ' style is thin 
village grain-store (with ' precious deposits ' ) guarded with 
fencing, which was translated into stone. This might account 
for the 'Yaksa' and 'Sri' sculptures in the early Stupa 
architecture, ($ri as a goddess is known to Sat, Bra.). 

4 Vide ante for explanation. 

5 Ev. X, 48, 7 (Nir. Ill, 10); of. Av. XI, 3, 9; VIII, 6, IS; Mait. 

cam. u., y, o, 

7 Rv. II, 14, U (Sayana). 

8 Vide V.I., I, 39-40. " 

9 In Ev, and onwards j vide V.I., I, 177 

10 Gf. ' kucakra ,' (so taken by Zimmer ; Alt. Leb. 157) ; Rv. X, 102, 

^"TT ?n dffl ,^ ^ EV * VII 36 ' 3 ? IX > 97 4 J *>^ ill < Suda-dohftS/ 

VIII, 69, 3) may be = well. ' Parsu * in Rv. I, 105, 8; X, 33, 
4} may mean the masonry sides of a ' kupa/ "with metal fittings, 
or ribbed, so as to resemble sickles. Similar ancient burnt clay 

g the sides of wells ma y be Been lin 

11 irr^af on. (in Rv '' =water -P i P es > an <* dug out water channels for 

12 Or 'wta.' Av. X, 3, 7; Taitt. Bra. I, 6, 9, 1. 

M w l/fl 1 ^* * **y%-wv* 'tj .. Jk _.. .. " .. * ^ . * Ji * " 


15 It scons that purchase of such a house with an ' udara ' of 


treasures" is referred to in Av. IX 3 3, 15, 

( 29 ) 

positions, vertical, horizontal and slanting. 1 (2) 'Vamfe's: 
entire bamboos, probably used mainly for the framework of the 
roofing, the central horizontal bamboo, supported on the 
1 stlrana ' or main pillar, being pre-eminently the ' vamsa.' 
(3) * Aksu * : either, the wicker-work or split-bamboo lining, 
over which the thatch was laid, 3 and to which the description 
of * thousand-eyed ' could aptly be applied ; or, a net, spread 
over the 4 visuvant,' to keep the straw-bundles of the thatch 
intact during stormy weather. 4 (4) ' Palada ' 5 and 4 trna ' : 
bundles of hay, straw, or long reedy grass, for the * chadis ' 
(thatch) , and probably for filling in off lining the walls. (5) 

* Visuvant ' : the ridge on the top of the roofing, looking like 
parted hair. 6 (6) Various ' ties ' joining the parts together, 
which evidently refer to bamboo and cane or rope work 7 ; and 

* sikya ' 8 : suspensory arrangements (like slings of strong net, 

1 This is more probable than bamboo posts and props, as * vamsa ' is 

separately mentioned; so also in Ev. I, 59, 1 and- IT* 5, 1, 
'upamit'spillar, probably of timber. Cf. the similar feature 
, in the Bengal f atcala. ' 

2 This term became early a technical one, denoting the main beam or 

ridge of any structure : e.g. the architectural sections of some 
Puranas know of the ' vamsa * of a fort or palace, where it 
cannot mean bamboo. Cf . the sense of * Jbeam ' in gankh. Aran. 
VIII, 1 and in ' gala- vamsa,' : Ait. Aran. HI, 2, 1. 

3 This is better than * thatched covering ' or * pole with countless 

holes J ; it corresponds to the ' catai * and ' jali ' of modern 
structures j cf . the current description ' cokh-cokh ' (* with many 
eyes ' ) of such wicker- work linings. 

1 ' Thousand-eyed * would apply equally to such covering net, which 

may have been of ropes or split-cane ; this sense is perhaps 

better, as ' aksn ' is said to the stretched as ' opasa ' on the 

* vis. u van V so that the net would correspond to *tbe finer net 

used to hold 'together the coiffure and stray curls- 

"5 With * palada ' and the cognate forms * palali ( = ' yava * -straw : 
Av. II, 8, 3), 'palava* (Av. XII, 3, 19; Jaim. Upan. Bra. I. 
54, 1), and ' palala ' (= straw: Kaus. Sut. LXXX, 27), may 
be compared the Eastern vernacular ( powal,' also a term specially 
used in house-building. A long grass,, * sirkl ' is still used in 
N. Bihar for such protective linings. 

6 As it actually does even now, the cut ends of the bundles of hay 
along the top being turred inside down and bound, so that the 
loose e-n<ds fall on either side. For the simile, cf . the * aku * 
spread over the roof is like an ' opasa * (woman's coiffure) ; the 
houso itself is likened to a ' vadhu * (and carried like her, on 
waggons probably, when dismantled) ; it (i.e. its spirit) is 
addressed as * manasya patni,* and is * clothed * in grass, etc, ; 
and the wife enters the new housei first, (So also* in subsequent 
thought the wife is * grfrfru ' as well as ' grha.* Cf. * the neat 
upon nest and vessel upon vessel ' of the Av. verse ins connection 
with this house). 

fl E.g. * samdamsa/ * pranaha/ etc. 

8 The modern ' sike ' (Beng,), suspended from the. roof to hold vessels 
and gourds, etc. ; they are sometimes made of woven cane and 
ornamental /designs. This may well have b'een the origin of the 
e ornamental hangings ' of later classical styles, as illustrated in 
Ajanta cave temples,* just as the whole of this type of dwellmg- 
honsV^s the source of manv later stone-architectural features. 
* Griffith's Ajanta Plates, No. 6, 10, 13; and Konow's Karpuramafi- 
jarl, n. p. 289, referred to in Whitney and Lamman ; Av. p. 536, 
where 'sSikya* is taken as = such 'ornamental hangings/ 


etc.). (7) ' Ita ',: which miwt be fine clay or unbaked bricks, 
rather than '''reedwork,' 1 used to finish off the \valls, or 
floor or basement, (8) Several Ride-rooms with a central hall 2 
(as indicated by ' paksas, 1 ' TviugK,' 3 * aguisftla/ the hall of 

With this * ita/ cf. 'the Eastern vcrn. forms ' lit ' (W, Bong.), 
Mta 1 (E.* Beng, and Bihar, etc.), ' itawa ' (8. Bihar and Oh. 
NSgpur). That 'it' originally meant 'clay/ is shown by the 
expressions c kafica" it ' and 'pukka it' ( -- 'pakva'}, ancMlje 
term ' etel ' or * aitla ' used of fine river clay, suitable for bricks. 
Unburnt bricks and m such clay are still used to fimah .and line 
the reed-walls or wattle. * Ita ' occurs in only another passage 
of Av. VI, 14 ; 3.* In both the Av. passages, ' ita '^ clay (or 
onbaked brick in IX, 3) suits better than the usual rendering 
of bulrush or reeds; in VI, 14, 3, U would mean the river-clay 
or silt washed away every year, and in the other passage it 
would mean the clay-plastering or ' kaftca-it '-facing, which had 
to be dismantled while the doors were taken off (of. the 
application of the verse in Kaus. Sut., 66 3 24). It is evident 
that this *' ita ' ' it/ etc.) has been Sanskrit ised into * istaka * 
by analogy; the original word seems to have been pre- Aryan. 
with an r or 1 associated with the t, the relic of which may be 
seen in 'etel' and 'aitla' (and- place-names like ' Ttli ' (Beng.) 
or f Itarsi ' (C, P.) ). So also, in Tamil (in the mod. form of 
which '| J is pronounced '4'), ' it(<J)a/t means to 'dig or dig 
out'; and ' ita-ppu * and ' ita- van ' (with which cf, ? itawa ' 
above, the plaoe-names 'Itawa' (south U. P.) and < IfjUwu, * 
(Travancore), and Tamil ( ddam '=site, house) mean 'clod of 
earth/ The Tamil for brick is ( ittika ' ; probably this is the 
original of ' istaka ' ; cf . the curious question in Sat. Bra. X, 6, 
1, , as to the (fern.) form 'istaka/ and its fanciful answer; 
the real explanation is the original Dravidian form ' itfrika, * (or 
the like, the ending representing the Tarn, suffix ' vaka '). The 
use of ^ clay and bricks therefore would seem to have been a 
Dravidian feature (of the Oangetic country) early introduced 
amongst the Vedic Aryans. This is confirmed by the curious 
Atharvavedic invocation of the dwelling-house as * Icja ' in the 
marriage ritual (Av. XIV, 2, 19) j this < Ida ' of course corres- 
ponds to the Dravidian forms meaning dwelling-house, traces of 
which may be found in the very ancient place-names of Mitliila 
l^Mithi^ila), Kampilla (=Kampi+iUuj cf, Mavella), or (the 
city) Krmila ( =Ki'mi+ila) ; cf. < Vi-deha (gha) ' and 4 Vai-sal! ' 
Divide ante); cf. also. Tamil < illu '= house, 'i<}a '= royal seat, 

i^a-vaka' -principality, parish or abode, ida til 1 in the 

seat or homestead, etc. 

* The Egyedic proper name * Ita ' may well be derived front 'it/ 
to wander, and would properly apply to a ' yayavara * rai ; Rv. 

' '^ K5vya ' : Km? ' Br5 - VI1 ^ 4 > ^ G - ^' 

* nn f c *! d with ' 145=8100806, which involved digging 

- ^ ra ' =the waste earth thrown U P by such 

the detaiied 

isti * *e famous ' 

36 3. ^e-room waa probably the Sgwa : ' Av, IV.. 

( .31 ) 

the fire-alt ar, 1 k havirdhana, ' the (sacrificial) store-room, and 
* patnlnam sadana, ' women's apartments ( 'site and seat 5 } s 
and with a covered * verandah ' running all around the house/ 
at the four corners of which were four thick-set pillars, 
probably of clay and rubble, or bricks 3 ; altogether a ' brhac- 
chandas ' 4 house, on a large scale and of ample proportions, 
covered by a 'many-winged '5 roofing. The prominence of 
bamboo, wicker-work, straw, and various ' ties J in the con- 
struction, and other peculiarities noted above, 6 clearly point to 
the lower Gangetic origin of this style. 7 It is very remarkable 
that the Atharvaveda which describes it, is pre-eminently a 
book of the Angirasas, who are definitely located in and asso- 
ciated with the very same lower Gangetic provinces in Puranic 
tradition, 8 Thus " the Av. style of housing is Eastern 9 

1 This must have been the central room (cf. Agni as embryo within 

the many-winged house), to escape dangers of fire, and also the 
front room, which would be first entered (as Is shown by many 
incidental references, e.g., in the Epics) $ it was also the "Vedie 
sitting-room, from the connection of the fire-altar with the sabha. 
The * havirdhana ' would be either the adjoining back-room, or 
one of the smaller side-rooms, whew the * soma vehicle ' could be 
dragged up and housed. 

2 At least alone the front and back," if a 2-' winged ' house. 

* The ' verandah * and 4 thick pillars are, inferable from the description 

in Av, IX, 3j 17: "covered with *trna' and clothed 
in ' palada/ the * nivesani ' is like a she-elephant with feet '* ; 
here the reference is evidently to the elephant with its fringed 
trappings and stout thick set legs, always clay-covered and clay- 
hued; the pillars could not have been of timber, for heavy 
timber work is incompatible with the * bamboo * style 5 thus it- 
is better to take them as rubble or raw-brick pillars; they cannot 
have been the ' upamits,' for these along with the slanting beams 
and the resultant angles would be filled in by walls (wattle or 
clay), BO that they would be indistinguishable as four thick 
iegsj thus these Megs" were independent corner-pillars, which, 
being under the same thatch (=the elephant's body, whose very 
curve of the back is like the ridge of such a house), necessarily 
implies the ' verandah * border, a characteristic feature of the 
Gangetic style, 

4 ' Chandas ' here may be compared with ' chaud ' (Beng.), a parallel 

structural term, implying the ideas of proportion, scale, or 
measurement, which is* also the sense of * chandas ' as applied 
to prosody; besides, as grains and cattle, as well as men are 
included within this * bvtiacchandas ' house, * proportions * would 
suit better than ' roof,*" cf . al&o the ' atichandas ' and other 
* chandas' bracks of Yv. altar construction. 

5 2, 4, 6, 8, or 10 ( wings ' (implying aa many s rooms *) ; cf. the 

division, of structures in Bengal according to * roofing * : * do-cala * 
(a poor man's house); ' can-cala (a thriving villager's house), 
,and ' ai-calS ' (a big open hall, used as the village ' sabha *). 

6 Vide foot-notes above. 

7 It is not necessary to go to the Nilgiri To<Jas for the vedic type 

of dwelling-house or for the originals of cave architecture (cf. 
V.I., I, 231). 

8 The significance of this point has been illustrated in several other 

sections q this dissertation. 

* This may be the significance of the 'first homage to the greatness 

of the house ' being paid from the * praeya, ' quarter (Av, TTC 
3, 29). ' ; 


groups. 2 Accordingly, references to some of its features are 
not rare in other parts of the Vedic literature aw well, 5 

But this type of dwelling-houses cannot have been the 
only one in the Vedic ages, and other varieties mut have 
developed according to regional conditions, etc. This is in- 
dicated by references to the use of materials other than the 
characteristically deltaic, for various structural purpose, e.g., 
of timber, burnt bricks, stone or metal ; and by mentions of 
other * parts * of houses, not named or prominent in the sketch 
of the above type. Thus doors and pillars do not form a 
special feature in this, but they are very frequently referred 
to in Vedic literature, and lead to various figurative uses, 
" Dor/ 4 the earlier and commoner word for door,? hus an 

* According to Chinese accounts (Ssu-ma-chien's Hist-, of ChSng-kien'a 

career and embassies), bamboos were imported from the Ciang^tic 

Delta as far up as the outlying North Western region of Ttt-hia 

(Bactria), as early as the 2nd century B.C., when it was regarded 

as a very ancient trade. The 'source of this overland and mari- 

time trade in bamboos, etc. was S- China and adjacent X'ndo- 

China. ^ (This agrees with the affinity between East Indian and 

Indo-Chinese types of house-structure, and the fact that Further- 

Indian bamboos '(being more solid) are still used in 

structures in some parts of Bengal) 
2 Vide Pargiter ; AIHT., p. 2193. 

E.g.~ (i VamsV : bamboo rafters or beam (Rv. I, 10, 1; Mait, Bom. 
IV, 8, 10, etc.). 'Aksu': (Ev. I, 180, 5 (prob.) ). * Sikya ' : 
(apart from Av, XIII, 4, 8, in) Taitt. Sam. V, 2, 4, 2, 3 ; 6, 9, 
1; etc.; Sat. Bra. V, 5, 4, 28 j VI, 7, 1, 16. Also 'chadia' 
(covering) and * upamit/ -rather more general terms, not 
restrictabl to the above type, {' Chadia ' ; Taitt. Sam. VI, 2 9, 4 ; 
10, 5, 7; Vaj. Sam. V, 28; Ait. Bra. I, 29, gat. Br&. Ill, 5, 3, 
9, etc. Cf. * chetdis s of bridal wagon : Kv. 86, 10- ' Upamit ' : 
Ev. I, 59, 1; IV, 5, 1). But ita' and ' pakda,' etc., do not 
occur elsewhere. ' Ata ' may be a primitive Aryan word, but 

*u e l? ct - that '** is " sed . of " the do rs of the sk y" 8nows tha* 
the Vedic poet had in mind not a rectangular timber framework 

for the wooden door, but rather % vaulted or arched framework 
af bamboo (cf. the style represented at the entrances to caves 
and cave temples), such as would properly belong to the above 
style g ate: ' Bv. I, 56, 5} 113, 14; HI, 43, 6; IX, 5, Sj Vaja. 
bam- XXIX. 5). The door-fittings indicated by * syuman ' 
r door-strap '; Ev. Ill, 6, 1, 4), and * dvara-pidhana * '( ' door- 
binder > : Sat, Bra. XI, 1, 1, 1), are referable to the same fltyle, 
vhrie 'argala 1 and ' isiH ' ('bar and pin' of cowmen: J9aAkh. 
Aran. II, 6) would rather belong to timber structures; (cf. 
vraja constructed of ' a^vattha ' wood). The frequenter uso 
of - *pL? 7 e plural to designate the house (vide the many 
rets, m Y.I. I, 229) shows that a number of rooms was a 
common feature; this may have been due to the early adoption 

/L * * f &ty e of houB e-lraildmjr with a number of * pake's 
(two to ten), r f 

?2- - 1 ', 6 ?' J; P- 4; 121, 4: 188, 5; etc. - 

" -' 13 !. 6 --' 4T- T 11 , 1 ' 3 ' 22 ' X1V '' 1. 63 ' VMa. Sam. 
. Bra, XI, 1, 1, 2; etc.- f dvark * ^ ATT Y 

; and ' 

( 33 ) 

implied sense of the. whole bouse, 1 and ' durya * (doorposts), 2 
' duryona ' 3 and ' durona,' 4 all signify the house itself; such 

use is an indication, that much was tfiought of the timber door- 

r*- J 

way, OQ which was probably lavished all tfye skill of the Vedic 
carpenter and carver. 5 * Skambha, 1 pillar (of timber), 6 is 
often used figuratively ; the somewhat later * stambha J? was 
probably sometimes a brick or stone'- one ; ' s>thann * s (the 
prominence of which is indicated in the use of sablia- 
sthanii ') and * Bthuna ' 9 are other quite common and 
early names for pillars (of houses or other structures^ , 
made of timber as well as other materials 10 * and the 
* sthuna-rajaj 11 must belong to a biggish complicated structure. 
Smaller timber po,sts were ' svaru 's 12 and * yupa 's, 13 used as 

1 Thus * dur-ya ' (In masc. pi )=* belonging to the door, or to the 

house': 1 Rv. 1, 91, 19; X, 40, 12; Taitt. Sam. 1, 6, 3, 1; Taj. 
Sam. I, 11. 

2 (In fern, pi.) ' durya ' = dwelling or doorposts: Rv. IV, 1, 9. 18; 

2, 12; VII, 1, 11. 

3 R V . I, 174, 7; V. 29, 10; 32, 8- 

4 Rv. Ill, 1, IB; 25, 5; JV, 13, 1; V, 76, 4; etc. Av. TO, 17, 3; 

Vaj. Sam. XXXIII, 72, etc. 

5 For such skilled artisans, cf. * taksan's ; Rv. X, 86, 5; Av. XIX, 

49, 8; cf. Rv. I ? 161, 9; III, 60, 2; ' tv,a,?t? ' : Av. XII, 3, 33; 
also 'tastr' in Rv. (vide V.I., I, 302). These artisans could 
make decorated and inlaid (pis) bowls like the starry night 
(Av. XIX, 49, 8), or the lotus (the ' puskara ' bowl of ritual), 
and could produce * rupam sukrtam * (sculptured designs and 
friezes?) with their chisels, and bowls had such carvings in 
relief of gods, etc. (Av. XII, 3, 33)- Cf. ' priya tastani 
vi-akta ' of Rv. X, 86, 5. 

The * taksans ' are respectable in the Rv. but have become low castes 
in the Buddhistic" age (see V.I. ? II, 266); the best explanation 
would be that these "wood-carvers" naturally enough amal- 
gamated with the Magadhan indigenous ' ' stone-workers " (vide 
infra.), and though as a result the crafts were much improved 
e.g. by renderings of wood-work in atone (as in the Buddhistic 
period), the craftsmen themselves suffered in status. 

6 Rv. I, 34, 2 ; IV, 3, 5 ; that it was originally a timber pillar is shown 

by the vern. * khamba ' = specially an entire ' sala * trank; cf. 
the expr. * latha-khamba * (an arrangement for drawing well- 
water), where" * khamba ' has that sense; (it is to be noted 
that the later monoliths are also called Math's; e.g. * Jarasandha 
ki lath'.). 

7 Kath. Sam. XXX, 9; XXXI, 1; and often in Sutras. Fop the 

"implication of brick or stone material, cf. the vern. use of 
' tharn,' ' thamba,* as comp red with ' khamba.' 

8 Rv. X, 40, 13; Av- X, 4, 1; XIV. 2, 48 j XIX, 49 10; etc 

9 Rv. I, 59, 1; V, 45, 2; 62, 7; VIII, 17, 14. Av- XIV, 1, 63 5 Sat, 

Bra. XIV, 1, 3. 7: etc. 

10 Eg. ' ayalisthuna ' or the sthuna on the grave (Rv. X, 18, 13), 

which 'may have been of clay or brick. So also the ' sthung- 
raja ' may occasionally have been of bricks., etc. 

11 Sat. Bra. IIL, 1, 1, 11; 5, 1, 1. 

12 Rv. I, 92, 5^ 162, 9, III, 8, 6, etc; Av, JV, 24, 4; XII, 1, 13; 


13 Rv. I, 51, 14. 

( 34 ) , 

.door-frames, etc., and ' methi 's 1 pc/sts for palisades. Apart 
from these varieties, used chiefly in houses, other pillars of 
different uses are indicated by ' sanku ' 2 (of timber as well 
as stone) and ' drupada/ 3 with which latter may be compared 

* ekambha ' and * vanaspati ' 4 (a pole or pillar,, evidently a 
dressed and entire pine or ' sala ' trunk). This great variety 
of names for pillars and posts, and the importance of these 
and doors, shows that they were a marked feature of .at least 
one other type of house-building. Thus, as compared with the 

* Deltaic/ there would seem to have existed, a Middle- 
Himalayan (and submontane) style also, characterised by 
skilled, heavy and prof nee timber-work : of which, again, the 
later and modern parallel is equally striking. 5 

To this timber architecture would naturally 6 belong the 
references to the use of metals in house-construction, such as 
the * ayahsthuna 's 7 (copper, bronze or iron pillars) and 
e parigha's 8 (metal bolts); and they must have been very well- 
known and prominent features to be used early as proper 
names. 9 There is no improbability involved iu this, aw in the 
early Vedic age 4 ayas ' was widely used/ and smelting 11 and 

1 With variants * m-edhi/ 'methi,' or * meibi ' : Av. VIII, 5, 20: 

XIV 1, 40; Taitt, Sam. VI, 2, 9, 4"; Kath. Sam. XXXV, 8; 
Ait, Bra I } 29, 22,- gat. Bra. Ill, 5, 3, 21; Paflc. Bra. X1I1,' 

3m 5 Ja i m ' Br ' T > 19 > isfor use ill palisades, cf. Bv". 
VlULj o5j 5. 

2 * Sanku '"usually = wooden post, peg, or even pin- (in Kv. and Bra.s- 

v ''> 1(I > 3 49) S but a stone pillar, in ' vrtra-s$imkitt> ' : Sat. 

2l 5> 15> and Scholia8t *" 

3 Bv .!,_ 24, 13; IV, 32, 23; Av. VI, 63, 3; 115, 2; XIX, 47, 9; 
Saip. XX 20; for use as posts for victims and offenders : 

' ' ' ad Av. XIX, 47, 9 above, and Av- XIX, 

Rv - 

5 Thus the rich carved wood-work and timber structures of the lower 

V^Ss PeS ? f ^! e J . Sonthein Himalayas (from Kaimir and 
Yamuna sources to Eastern Nepal, and the submontane plains 
Sbah , ara P ur and other districts) have all along been 

of I" 086 Pte. (Cf . the place namea : 

and ' Kath-godam ') 


in metal work is found side by side With that in 
,w * - ,u- v ab v r S>n (cf. the well-known artistic 
P, P w .? o hlS lme of Ne *&> Moradahad, etc-)- 

JXV, V Uu f , Q 

L '' II 4 24j T}*- 10 ' 15; ' and often later '' 

' : . a *\ J? r5 / ,^I> ^ 2, 17 (the name belongs 'to an 

ft 1i ' ' i_ O . 

* / '8 name in the PiuiSinip 
IQ T - v- ,), step No. 42. 

m,s* = smelter. 'dhmatarV smelting furnace: Bv V 9 S- 
tog: *r. IV, 2, 17; ltteieltowa : Hv IX 112 2-' 
ino- ores (aiman) : Sat, Bra. VI, 1, 3, 5, ' , ' ' 

(, 35 ) 

beaten 1 k ay as ' 2 are referred to ; the * ayahsthuna's and 
'ayasi pur's would thus imply the strengthening of" tina&er pil- 
lars, palisades or walls, by copper or s.tee'l 3 -platicg and sundry > 
metal fittings. This would constitute a necessary earlier stage 
of architecture to account for the elaborate gold-plated and 
inlaid timber-pillars of the fourth century Mauryan palace. 

The first explicit mention of the use oi burnt bricks 
( s pakva J ) for structural purposes occurs rather late, in the 
Satapatha 4 (6th 7th cent. B.C.); but even there, this 4 pak?a' 
and the * istaka/ which is used throughout, are taken as 
identical ; and as the reference is to the building of sacrificial 
fire^tlta-rs/it is clear that this use of * burnt * bricks was more 
or less traditional, 5 and not a recent innovation 6 ; besides, various 
well-known personages are stated to have erected such fire- 
altars, 7 some" of whom can be approximately fixed in time with 
the help of * traditional ' chronology : so that such construc- 
tions would go,, back to the earlier Vedic period. 8 * Istaka * is 
indeed the traditional material 9 for building the fire-altar even 
in the Yv. Samhitas 10 ; and though not specially called 'burnt-/ 
these bricks were almost certainly so : for it is often stated 

1 For ' soma ' vessels : Ev. IX, 1, 2. 

2 The use of * sheet ' iron is more probable than cast iron, though 

the antiquity of ore-smelting (probably pre- Aryan) and the 
quite early occurrence (cir. 300 A.D.) of massive and highly 
finished foundry products, may indicate an earlier long standing 
use of cast iron posts and rods for structural purposes. 

3 Indian steel was well-known in the far Western countries in the 

6th and 5th cents, B.C., and was as much prized by the Greeks 
in the 4th as tributes of precicws gems. It is quite likely, 
therefore, that * steel ' should have been variously used for strength- 
ening defences within India itself, before its fame spread abroad. 

4 Sat. Bra. VI, 1, 2, 22; VII, 2, 1, 7; in the former passage it is 

said that the ' pakva ' is called ' istaka * because it is.Jis.ta/ 
offered to the fire (the derivation being a late etymological 
fiction ; cf . the fanciful explanation of the form * istaka ' rather 
than ( a * or ' am * : ibid, X, 5, 1, 5 3 also vide ante, re 
' ita *) 5 in the latter, a special * black * ' pakva ' is made by 
baking the brick in * rice-husk ' fire. ' Pakva ' in %., AV., 
and Bra., means simply ' baked/ or ' cooked food ' (vide V.I-, 
S.Y.); in Rv. VI, 63, 9, however, the sense of baked bricks, or 
a ( house of baked bricks * (a ' pucca ' house), may suit quite 
well (as horses, slaves, chariots, etc., are given away by certain 
Kings, Puraya, etc., to the priest, along with pakva*). 
(N.B The substantives are all understood in this passage-) 

5 I.e., representing Yajurvedic (Vajasaneya) tradition of a much 

earlier age. 

b Cf. the conservatism of the gatapatba regarding proposed changes 
in Rgvedic texts (and to a less extent in Yv. texts). 

7 E.g. Tu'ra-Kavaseya : Sat. Bra. IX, 5, 2, 15; gyap&rna-Sayakayana : 

ibid. VI, 2, 1, 39 ; IX, 5, 2, 1. , 

8 E.g. Tura-Kavaseya.. temp. Janamejaya-Pariksita I, cir. 20 steps 

above the close of the Rgvedic period. 

9 Cf. ' istaka-cit ' : Taitt. Sam. I, 5, 8, etc. 

10 E.g. the IVth .and Vth books of the Taitt. Sam. (mantras and 
explanatory matters re ' agnicayana '). The details regarding 
altar construction in these are practically the same as in the 
gat. Bra., thus showing that the use of bricks was traditional 
and almost co-existent with brahmanism. 

{ 36 ) 

there by way of explanation, that bricks were invented 
apparently by the Angirasas) to save Earth from bein^ 
excessively burnt 1 by the sacrificial fire 2 ; their supporting 
strength 3 and capacity of resisting the waters 4 are often 
specified; and amongst the many types of bricks used, were 
the ' svaymatrnna >5 or * naturally perforated ' bricks, and 
' bricks of a-U colours/ 6 the former being >a characteristic 
product of the kiln, 7 and the latter probably referring either 
to the various shades of red in the brick-piles, or to enamelled 
bricks 8 ; while mortar Cpurisa') that could be compared, to 
flesh adhering to bones, 9 had probably an admixture of 
pounded red bricks. The art of brick-laying was an old and 
developed one ia the Yajurvedic age, judging from the great 
variety of c-aines aud forms of the alter-bricks, amongst which 
may be mentioned the ' circular bricks ' (' mandalestaka *) 10 
the * earless ' or corner-less bevelled bricks ( " vikirnl * ) , u 
the * crest * or conical bricks ( ' coda ' ) , 12 the * gold-headed ' 
bricks (' vamabhrt '), 13 the shaped ' pot '-bricks C kumbhes- 
taka T }, 14 and other bricks with various linear markings 15 and 
of different sizes. 16 Mortar (of mud and rubble, sand or 

1 Taitt. Bam. V, 2, 10; 5, 2, etc. 

2 It is indeed only natural that the use of the baked' bricks should 

have early suggested itself for sacrificial structures, for the 
properties of burnt clay would be evident to any fire- 
worshipper; besides, with the growing ritual importance and 
significance of the altar, square or rectangular bricks must have 
been invented or adapted, and these, if unbaked at first would 
soon suggest the burnt brick. 

3 E-g the 'asadha brick of thousandfold strength,: Taitt. Sam- IV 

TIT it ' c he bnck that q ua keth nob' (' svayainatrnnl ) : ibid' 
IV, 3, 6; y, 6, 2; c Brhaspati saw in bricks the support of 
sacrifice : ibid. V, 3, 5 ; ' brick-altar ' representing ' the fi*m 
earth m the midst of waters : ibid. V, 6 4- ' 

E.g Taitt Sam., V, 6, 4 (in n. 3 above) : ' bricks keeping the altar 

^ SW * ' "- V, 3, 10- 

7 This is 

s is called ]hamS in vern.. meaning perforated (cf. the cognate 
words lhanjhra/ and ' jhanjhari, ' of same signification). 
Enamelled earthenware and tiles have long been a speciality of the 
9 T *ttT Ml ^ r dle 2 Ga ^ etic Districts (Eastern U. P.). 

n J ai . tL Sai?1 * ^ ^ 5; V, 3, 9; etc. 

I w *B* JLj_ ft ~lt"f' ^W ** . , 

topmost, and 

- -- "' *. ticii\A (jj_- vauLi-aibtitir ~ oncKS ; appar 
of ornajnentation). 

12 Tatff Saw TT7 ^ 7 /!_ _!__ -. i-, t ., m ^ 

! 3, 7; etc. 

p Qnjasi s] B -- at the 

" Taitt. Sam. V, fi, l ; etc. 

'5 Taitt. Sam. V, 2, 3; 2, 10 (cf. S.B.B. xliii, p. 21 , n 1 rr lines 

on square and rectangular bricks); of ' the various tv of 


of course involy^ different sizes, 


pounded bricks) was freely used 1 in " making bricks firm/' 2 
cementing successive layers of bricks, 3 and in plastering over 4 ; 
such adhesive plasters must have been essential in the con- 
struction of the alternative forms of the altar, 5 like the 

* bird '-styles (representing the ' syena/ * kanka * or ' alaja ') , 
or the * bowl ' or ' granary ' (' drona ') , 4 chariot-wheel/ 

* circle/ * cemetery ' (smasana) , and ' triangle * models. 
Large numbers 6 of bricks were used ' for these altars : the 
measurements of one altar is given as 36 feet along the centre, 
E. to YV., and 80 and k 24 feet across at the back and front. 
respectively, and it is said the outer limits of the^measurements 
of the altar depends on what area the builder thinks he could 
very well use 7 ; the first, second and third pilings are to be 
made of one, two and three thousand bricks respectively 8 ; and 
the bricks ready before an altar-builder (who wishes those 
became his cows) are roundly estimated 9 at hundreds of 
thousands. The rites performed on leaving a homestead, with 
a view to re-establishment elsewhere, *show that in the ordinary 
household also the altar was brick-built, and apparently these 
bricks were dismantled, carried to, and refitted in the new 
4 vastu.' 10 . It would be extraordinary if bricks were not used 
for the secular house-buildings as well, while altars (house- 
hold or .special) and cemeteries 11 were brick-built. It is remark- 
able that throughout the "brick" mantras, reference is made 
to " the manner in which Angiras placed 12 the bricks firmly," 13 
or invented them, 14 or used them for better building of the 

1 E.g. in Taitt, Sam. I, 2, 12; II, 6 y 4. 

2 E.g. in Taitt. Sam V, 2. 3. 

3 E.g. in Taitt. Sa,m. V, 6, 10. Cf. 'seasonal' bricks being "the 

internal cement of the layers' 7 ; ibid. V, 4, 2. 

4 ' Just as bone is covered with flesh ' ; Taitfc. Sara. V, 2, 3. 

5 Taitt. Sam. V, 4, 11. (The structural peculiarities ol some of these 

types," according to the Sutra comment, were; a round-topped 
block (the head) for the 'kanka'; curved * wings/ for the 
'syena ' ; ' caturasra * or ' parimandala ' (square or round) 
blocks for the ' granary * (< drona ') j and square OT round form 
for the 'smasana'). 

6 Cf , similar large numbers in the Sat. Bra. ' agnicayana directions : 

e.g. 756 bricks: gat. Bra. X, 5, 4, 5. 

7 Taitt. Sam, VI, 2, 4. 

8 Taitt. Sam. V, 6, 7; (the height of altars being up to knee, navel, 

and neck, respectively). 

9 Taitt. Sana. IV, 4, 11 ("these bricks hundred hundred 

thousand millions"). Of. ibid. V, 4, 2. 

10 Taitt. Sam. IV, 4, 10. Cf. the dismantling of i|a etc. and 

carrying of them in the Atharva-vedic mantras (vide ante-). 

11 The direction that brick-altars could be erected after the model of 

(round or square) smas'anas/ shown that these latter were also 
brick structures by the time oj 1 the TV. Samhitas. 

12 Taitt. Saro. IV, 2, 9; 4, 3; etc. ^ 

13 So also in Sat. Bra- the expression is repeated: .e.g. ^, 5 T 1, & 

(' sadan^ settling of the brick, may be the original for the 
vern. phrase 'it sajana '). * 

" Taitt. Sam. V, 5, 2; ci V, 2, 10; so also bricks are 

have been " fashioned by the toils of seers like metres : v, 

( 38 ) 

fire-altar 1 ; sometimes Brhaspati (also tin Angirasa) is 
Introduced; 2 and the brick (Tstaka) Is addressed and wor- 
shipped as a goddess deyi ').* All this is strikingly similar 
to the expressions and notions of the architectural sections of 
the Puranas^ where the laying of bricks and other stages of 
house-construction are accompanied by references to the 
Angirasas and their deified ' daughters.' In view of what has 
already been said about the ordinary ' brahmai.iic ' dwelling- 
houses of the Gangetic type (as described in the ' Anpirasn 
Veda ') and the use of clay and unburnt bricks (* ita ') in 
them, 5 and of the fact that the dwelling-house is addressed in 
the same Veda as ' Ida * (which also is evidently connected 
with the Dravidian roots and words meaning digging, bricks, 
and house) , & the inference__becomes irresistible, that this 
consistent association of the Angirasas with the invention and 
use of * istaka/ in Vedic as well as Puranic tradition, is but 
another 6 indication of a fact of cultural history, that the 
civilization of ' brahnianism,' with its sacrificial cult and 
symbolism, its building activities and material achievements 
and equipments, was originally Gangetic, Eastern and non- 
Aryan. In any case, it is quite clear that a third structural 
style, characterized by the use of clay, plasters, and bricks, 
dried or baked (of diverse moulds and probably even enamelled 
sometimes) was already in existence In the 10th century B.C., 7 
being referred to in the Brahmanas and the later Samhitas, 
and is implied for the earlier Egvedic period 8 ; and "here too , 
the conclusion agrees with the regional indications of the 
references : for this style can only have arisen in the riparian 
districts along the north of the Ganges (middle and lower). 

It is in the Satapatha again, that the first clear mention 
of stone structures of a recognizable type is made, but in a 
way that would indicate a well-formed, distinct and traditional 

1 Taitt. Sam. V, 3, 5. 

2 Vide n 14, page 37. (Some special forms of altar-bricks or manner 

of laying are associated with Visvamitva and his contemporaries: 
this may indicate the taking up of brahmanical sacrificial cult 
and connected brick-building by the Ailas in that period). 

3 lain. fcam. IV, 2, 9; cf. the house goddess c manasya patni ' fixed 

by the gods in the beginning (Av. Ill, 12, 5), and Brhaspati 
first putting together the house (Av. IX, 3, 2-3). (Vide p 31, 
and notes 7 and 8 IE it.) 

tile / vSstu - m ^ a ' sections of Agni, Garuda, and Matsya, 

5 v ide ante. ' ^ 

7 fL^nf!.f n<li< if* i M fa J the , evi , dence about '^rniture' etc., infra. 

H a h ' S ( hly t*"** ar t of making bricks, glass and 
pottery of various sorts, dating from cir. 3000 BC., have 

Trthn^-/ 66611 ^," 1 he PP er and lwor Indus 
to IL 2 & ^ 1 ^? bable that the literar y evide ^ with 

i0r s*r e ificS7 C R ^% and W T 1 ^* been kn t ^e 

,=s -as: s 

m y 

style. Its remarks on the erection of 'smasanas ' (over burnt 
or buried bodies) .are significant. They show a marked 
difference in the contemporary modes of building 1 these funeral 
and memorial structure**. The ' Praeya ' mode of erecting 
tombs is strongly disapproved 2 (from the point of view of the 
Kuru-Paflcala and Videha 3 brahman), Apart from minor 
difference within the approved range as regards special forms 
for the several orders, 4 the structural type that is regarded as 
unorthodox is described clearly as round and dome-shaped 
(' parimaiiujala ') 5 ; that whereby the Easterners make the 
' tmiasana"'* 4 ' separate from the earth " unlike good people, 6 
is described by the usual Vedic word for a large hemispherical 
bowl, 7 * camu> T which must here refer to something like a wait 
or dome of solid stone or bricks 8 ; the structure is then 
* enclosed by an indefinite number of enclosing stones ' 9 ; and 

1 fiat. Bra,. XIII, 8, 2, 1. 

3 Ibid. XIII, 8, 1, 5; 2, 1; of. IX, 5, 1, 64. 

3 T4w preference for nprth-indinad and saline soil poiuia to a Videna 

origin of these views. 

4 Ibid. Xllli, 8, 2, 6-12; 3, 11. 

5 Ibid. XIII, 8, 1. 

materia (without any preference for one 


i ( pUed m g o U nd is fi/ally beaten, v ('kuay 
a/ with a number o &la' or 

round -forms in the 

threcognn roun or 

smasanal in the later Sanihitas (vide p 37, n. 5 and p. 4B) ^t 

xvhich ultimately found expression in the Buddhistic reionna 

* ,* 

H.J, i^ j^ JuJlfcT^' nk* l v nF 1 ' 1 *' -i ^ _* . i^t 1 jb i^l f\ *% JL 

10, it is evidently filS and utaka that are placed and 
Fofwition of round, forms,, of. also Taiti .Sam IV ; 3 

** * ^O . a 1 *_^__ ^_ _ .t<wi.p*I/\\a 1 */ Li. 411 I DLriiCjl 

3 d - 

(noted above). 

( 40 ) 

stones are used instead of the square bricks in the case of 
fire-worshippers. 1 The orthodox style of ' srnasana * is stated 
to be square or quadrilateral, 2 ' not separate from the earth/ 3 
(i.e., not prominent and towering 4 like the banned type, and of 

* earth and earthen ' materials, clay and bricks, as opposed 
to stone), and bricks one foot square are used in its 

construction 5 ; and a memorial mound like .a fire-altar 6 is 
prescribed 7 for builders of the same. 8 It is evident that the 
former is the prototype of the Buddhistic, Eastern and heretical, 

* stupa' architecture of the very next epoch, and through it 

of the ' Saiva ' temple styles of subsequent ages 9 ; and that the 
latter is a specially ' brahnaanical * style, associated with. 
sacrificial altars 10 and the middle Gangetic country, and thus 
with bricks 10 and rectilineal figures, strikingly paralleled by 
the similar sacrificial and geometric style of squares and bricks 
in .ancient Babylonia, and represented recognizably in some 
later forms of ' brahmaiiical ' temple architecture. 11 

1 gat. Bra. XIH, 8, 4, II. 

2 Ibid. XIII, 8, 1, Iff. 

3 Ibid. XIII, 8, 2, 1. 

4 The Satapatha insists repeatedly on the * smasaua * being .{lot 1 oo 

large or high: e.g. XIII, 8, 1, 18 (an ordinary altar's sisse) ; 
8, 2, 6 12 (generally and preferably to be knee-high, though 
structures as high as the thigh, hip, mouth and upstretched 
arm, might be allowed for vaisyas, women, brahmans and 
ksatriyas, respectively; note K^atriya superiority). 

5 Ibid. XIII, 8 3 4, 11 ; not l marked ' like altar-bricks. 

6 It is noteworthy that about 3 centuries later, Alexander used fire- 

altars as * memorials/ apparently according to the Indian custom ; 
to impress the Indians he is said to have built on the Beas (ct 
the custom of building fire-altars on river-banks, indicated in 
Sat. Bra. and earlier as far back as the Rv.) stupendous and 
sculptured fire-altars of stone, which Candragupta later on 
utilized for sacrificial purposes. 

7 Sat. Bra, XIII, 8., 1, Iff; sometimes * without wings and tail,' i.e., 

in the form of a simple cubical altar, without the 3 adjacent 
cubes; the special recommendation of the Satapatha is an 
irregular quadrilateral with sides joining at S. shorter than those 

- ; > !? ut thls ma y refer to tne area enclosed by cords, within 
wnicn the altar-like smasana is raised. 

s This may imply that those, on the other hand, who built (and 

worshipped at) the round stupas, were similarly honoured by 

" * TS? mneral memonals. It may be noted here that worship 

of the funeral mound is implied in Av. XVIII, 4, 38 (it is 

2S; Jj!! 1 b C ? s on wors %pers), and that the 
previous Buddhas a l so had their stupal' ' 

8 ^DSL^ th p r Und I * ** h been designated 

( 41 ) 

The c Pracyas s referred to here cacnot be those deltaic and 
riparian Easterners, 1 to whom the Atharva-vedic style of house- 
building most be attributed ; the passages in the Satapatha 2 
may be taken to mean " the Asurya section of the Pracyas/' 
i.e., either the unorthodox Magadha'n Pracyas or the Pracyas 
who follow Astiri's tenets, the proto-Buddhistic creeds (the 
association ol' round stone structures with them, in the latter 
case, being historically sound) ; the very allusion to solid stone 
or brick vaults, stone enclosures, and stones as substitutes for 
bricks, shows that the region meant is Magadha, 3 known as 
PracI pre-eminently, in the 4th century B.C. [Magadha^ and 
Kikata are looked down upon in early as well as later Vedic 
literature 4 ; and it is precisely these regions 5 which have an 
ancient tradition of stone masonry and ware ; so also, when 
the cars of the Pracyas, the ' vipathas,' 6 are disapproved by 
Midlanders, 7 it is evidently the rough country of Eakata- 

1 Of VaisJali, Anga, Vanga, etc.; vide ante. 

2 Sal. Bra. XIII, 8, 1 ; 8, 2, 1. . 

3 It is noteworthy that so far the earliest known remains of vaulted 

and polished caves, of stone enclosures, walls or pillars, are in 
Magadha or of Magadhan origin. 

4 E.g. in Vaja. Sam. XXX, 5, 22; or Kat. Sr. Sub. XXII, 4, 22; vide 

also note 6 below; Cf, also the famous Rv. reference, "Kim 
to kvnvanti Kikatesu, etc." lu Puranic tradition (cf. Vayu. 
78, 2122) the land of Trisanku, bet, Kaikata and the 
MahanadI, is avoided by orthodox people. Kikata and Gaya 
arc almost identical in Va. 105112- So also, the benighted 
region where Trisanku is banished seems from epic indications 
as well (cf. e.g. all that is said about v ^5mitra, Matanga and 
his tirtha:-Mbh. I, 71, 2925-28 ^h Hanv V 717fL; III, 87 
8321 (in the East): III, 84, 8079; HI, 85, 8159; XIH 27-29 
(Gaya); XIII, 3, 189 (in the South); cf, Varaha. T and VIII 
(conn, with J\Iithil5 and Orissa) to have been no other than 
Kikata (cf. also popular traditions re Rhotesgarh and B. 
Karmanaia). And if the Kicakas of Mbh., whose country the 
Pandavas passed just before coming to Ekacakra, anxl who cremated 
their lusty chiefs with their women, are the same as the Kikatas 
(vide infra, sec. re widow-burning), -it is another trace (even) 
in the epic literature of the low estimation of these Pracyas. 
s Now represented by Gaya and Cunar ; also similar regions westwards 
along the Vindhyan borderland, Jubbulpur, Gwa hor and Jaipur 
ZrLenting ancient Cedi and Matsya, very closely connected 
with Magadha In the Puranic tradition. 
!>- Tt^-; -vvTr 1 (A verv old passage) ; Lat. Sr. out. viii, o ? 

5 }^*'^^^ 


reide in a Nisada settlement (S E of 

but there were 

III, 44), thus 
bches oi loresis m ws wcu v ***" *-. J - iJ -j / . 

that architectural styles must have been largely of 
' JBastern ' origin. 

( 42 ) 

Magadha 1 that is referred to.] ' Smasana 9 structures of the 
two types distinguished by the Satapatha were evidently 
known in the earlier Yajurvedic period. Thus a fire-altar and 
a ' smasana ' are similarly piled, so that the former has to be 
differentiated by burying a 4 living ' tortoise in it 2 ; again, 
certain altars are piled in the form of ' fimaaanas,' which, 
according to the Sucra comment on the directions, are of two 
well-known types, round or square, 3 just as the ' drona's or 
4 grain-stores,' which ateo supply the models for other types of 
altar, were round or square .structures. 4 In the Av. and Kv. 
also, it may be a round type of ' smasana ' that is set up, with 
f * Swell thou up (ucchancasva) ... let the earth remain 
swelling up ... let a thousand props support it " 5 ; while the 
funeral structure that is said to be ' cayanena citam * 6 is 
obviously of the same type as the square altar. Knowledge of 
big round structures like the ' stupas ' (or ' camu 's of the 
Pracyas) is suggested by the metaphorical use of ' caniva ' in 
Ev. to denote the vault of heaven placed on the earth 7 ; so 
also the 5gevdic use of the word * stupa ' itself clearly* B!IOV;S 
that it was a structural term as well : thus Agni on the altar 
"extends up to the sun's disc with ' stupa's of flames/ 8 and 
" Varuna upholds the ' stupa ' of light on the baseless 
firmament. 3 ' 9 

The Satapatha classifies ' smasana ' structures into the 
ordinary ' vastu ' or reliquary of bpnes, etc., ' grhaii 1 and 

i Where the ' sagad ' and the ' ekka ' are still characteristic 
conveyances evoking much comment (for a humourous satire cf. 
the mod. Beng. ballad " Vi ghore Vihare cadinu ekka," etc.). 
7 - e *,^ a<3 * is characterizec. by solid timber or stone wheels, 
* saJa '-timber body and a peculiar drowsy long-drawn squeak 
heard from great distance (cL ' s'akata J in the ' aranyaiii ' hymn 
in Ev.; cf. also the peculiar construction of the traditional 
toy-cart, * mirt-sakatika *) ; it is comparatively low- Wilt and drawn 
by buffaloes, and can be drawn over all sorts of rough tracks and 
^gions. The (one-hoFsed) ( ekka ' is probably alluded to in 
Vedic passages where conveyances with a single horse are 
deprecated; (generally, in contrast with those with 2 or more 
horses: Bv. X 151, 3; Taitt. Bra. I, 8, 2. 4; III, 8, 21, 3; 
|? c - Bg. XVI, 13, 12; XVni, 9, 7; Ait. 6ra. V, 30/6; Sai 
Bra. AMI 3, 3, 9j etc.; jpopr people content with one-horsed 
car: Bv X, 101, 11; -VI, 15, 19; Pane. Bra. XXI, 13, 8j etc.). 
lae 'ekka also is suited for rough country use, and might 
well be called * vipatha.' 6 

I l a S" l al|s ' I 3 % & 5 (ci >dlso ^ b - Sai ^- m * *> ^ smasana). 

3 Taitt. Sam. V, 4, 11, 3, 

4 Taitt, Sain. V, 4, 11 -, the Sutra ascription of * caturasra r and 

^ panman^ala styles to both funeral tumuli and grain-stores 
is mtejesting ; vide ante re connection between * etupas * and 

5 Av. XYIH, 3, 50-51 =Kv. X, 18, 11-12. 
Av. XTIII, 4, 37. 

T Bv. Ill, 55, 20. 

8 BT. VII, 2, L 

S Ev. I. 24, 7; * stupa 'in Av. VI, 60, 1, is used of the round coil 
? -A^ Hainan *s head; tne figurative us^here and elsewhere 
ivitte y.I., s.v. * stiipa ' ma y well be compared with the 
comparison of the homse-top wiil^ * opasa f and parting of the 

( '43 ) 

' prajfianam.' 1 The first is evidently the tumulus, round or 
square, which forms the subject of so much comment in that 
Brahmana. The term ' grhan ' used of a special type of 
' smasana ' is particularly interesting : properly it means a 
dwelling-house with many chambers ; applied to a ' smasana * 
it would signify that the funeral structure was either an actual 
house (mausoleum) with many rooms, erected over or beside 
the grave in memory of the deceased, and for the benefit of his 
soul dedicated to some religious order, or philanthropic use, 2 or 
that these ' grhan ' are the chambers and vaults of sub- 
terranean or rock-cut caves. 3 ' Grhan/ however, Is nothing 
new in the later Brahmana age, for the "Av. (as well as the 
Ev.) mentions it frequently*: thus referring to "a funeral 
structure it says, " let these ' grhasah ' be a refuge for him 
for ever " 5 ; elsewhere, tc make ye ' grha's for him according 

1 Sat. Bra, XIII, 8, 1; (cf. also cumm. on it). For l vastu ' in this 
' sense, cf. Kapila-vastu, where the sense must, be " the memorial 
'stuoa' of Kapila" rather than the 4 abode of Kapila,' as 
usually taken. Kapila lived in the middle of the 8th century 
B C. according to Puranic evidence (vide Pargiter: AIHT, pp. 
330332); hence, it is evident that the styles referred to in 
gat. Bra. were at least two centuries earlier than itself, 

. . 

important point. , , 

All this is characteristic of Buddhism in the very nest epocb, and 
traces of Buddhistic features can only Ue expected m the later 
Vedic literature. . 

This also would be a Buddhistic feature; relics were deposited m 
rock-cut caves in historical times ; in the Epic the rock-cut caves 
of Girivraja are used for condemned prisoners or human ^victims, 
and other caves are also said to foe similarly used; tho Epic also 
knows of ascetics in subterranean caves; the Barabar caves may 
have been intended as memorial ' smasanas ' of some Maurya 
emperors, presumably Asoka, etc. The Roman catacombs and 
Egyptian cave-graves offer instructive parallels. Another 
remarkably Magadhan and Buddhistic feature found in the 
e smasana'* of the Sata-patha is the regulation " let there be 
'citras' on the back oi: the ' Sjaasana,'" "for citras 'mean 
offspring," (The comm. takes it as=natural scenery; this is 
absurd, specially as natural scenery is suggested as an alternative 
in the following lines). In the case of the brick-built tumuli, 
these * citras ' would be ' paintings ' on_ suitable plaster, but in 
the case of the stone-built round * camu ' of the Easterners the 
most suitable ' citras J would be sculptured figures in relief ; tfte 
nature of these 'citras' is indicated, by the reason given; tne 
Bgures painted or carved were of women and children, and 
possibly couples of men and women. It is interesting to compare 
the account in the Epic of the representation of _ the_ fertility 
' gcddess Jara (or Jata; cf. the traditional village spirit, Jata-bu^i) 
on the palace walls 'of the King of Girivraja, of a plump woman 
with children all around, and also the panels of female figures, 
amorous couples, etc., in the later ' stiipa ' and ' vihara archi- 
tecture (cf- the Orissa temple sculptures). 

Besides, ' smassana ' and { sadman ' (house) are often spoken of^as 
parallel things: e.g. Av. V, 31, 8; X, 1, 18; so 'a so by burying 
a live tortoise an altar becomes a * vastavya and not a 
'smasana' (Taita. Sam. V, 2, 8, 5); (probably there is an 
implied pun on * vastu * here). 
Av. XVIII, 3, 51 = Ev. X, 18, 12. 

( 44 ) 

to his kindred "*; again, " as the 5 clans Cmanava 1 ) 
implanted a * harmya 5 for Yama, so I implant a * faarmya. ' 
that there may be many of me." 2 It would be too much of 
a forced explanation to take ' grhan ' as a metaphorical 
expression throughout , .specially beside the technical sense 3 
given to it by the Satapatha ; even in the Bgvedlc description 
of the grave as a ' mrnmaya grha '* into which one goey down, 
though there is an element of figure, yet the use oi: ' grha ' 
seems significant': it is possible that the phrase unconsciously 
refers to subterranean burial chambers or vaults. The 

* prajnanam ' of the Satapatha (beside the ' reliquary ' and 
the A chambers ') can only mean some sort of a memorial 
monument, like a pillar. A pillar (sthuna) indeed is set up 
on the Bgvedic grave- 5 (in the Av. also); "and a 'loga ' (pole) 
is erected after the earth is piled up (' ut-stabh ') from about 
the grave 6 ; and on the " sthuna ' ' maintained by the Fathers ' 

* Yama makes seats for the departed J? ; and it is probably such 
memorial pillars (on which the spirits ' sit ') that are referred 
to, where ' the bride-beholding fathers * are asked to be 
propitious to the brixie as the marriage-procession passes a 
cemetery. 8 These ' prajfiana ' pillars may have been of timber 
originally 9 ; but as bricks or stone came to be used for the 

1 Av. XVIII, 4, 37. The qualification ' according to his kindred, 

evidently means that the size and excellence of the * grM? "' 

depended on the number, position and means of 111 kinsmen 

of the deceased (cf. its exemplification in actual Buddhistic 
dedicatory structures) . 

2 Av. XVHI, 4, 55. The use of * harmya * is significant, as in early 

Vedic literature harmya ' has the sense of a big establishment, 
with many apartments and adjuncts, and is used also of kings' 
residences. The motive of building a funeral ' harmya ' as Riven 
above is noteworthy: it foreshadows the dedicatory buildiries 
and parts of them in the subsequent Buddhistic age. 
Ev VII 6 f 1 a g Structure involv ed in the city-name < 

5 Bv. X, IS, 3=Av. XVm, 3, 52. 

6 AT " ^Id 1 ^. L 52 > EV - X '- W i 3; 1 ga ' here is usu 

clod _ % but it seems m the next passage to be identified with 

iJrf52% a 3r P ^^we means a pole (stuck into the 
bottom of the waters, m marriage ritual; vide Kaus 75 14 



o AV TTV 9 7T -L^**"" F" i . , iKmour 01 a late king. 

tSlarfv . - P ? raS ?, * nd *Mi0idmg fathers* would be 

particularly appropriate if the reference were to sculptured 

., 3, 33) ; grave-posts with effigies and 

face Or eves ara nnf nnlrnrnirr. *mnnc. 


& " - v,AAb TlM'fcJ VDn*iVM 

are not unknown amongsi 
buried in hollowed^out tree-trunks 

, r III. 3. 70/ r.f. 

. 2/25, 

( 45 ) 

funeral tumuli, 1 these also would be of the same materials by 
aiid by; thus in the time of the Satapatha a stone-pillar 
(' sanku ') 2 was set up along with 3 timber ones at the four 
corners of the ' smasana.' 3 The Buddhistic monolithic pillar, 
erected beside the relic-stupas and on the 'highways and public 
thoroughfares, 4 is probably the developed form of such -memorial 
' sanku's ' and the civic and sacrificial Vedic ' drupadas * 5 
(symbolical of royal 6 and divine power 7 ) to which offenders and 
sacrificial victims were bound: as the symbolism of the 
1 smadana * structures developed with and under Buddhistic and 
protp-BudtMstic thought, 8 and as offenders ceased to be 
punished so brutally, and sacrifices fell into disuse, these 
* sanku ' and ' drupada ' pillars would be used for ethical 
purposes and ' dhamma ' edicts (just as the traditional royal 
hunt was transformed into missionary tours) . 9 This is 

* Apart from the clear instances of the use of s-tone for the * smasanas ' 
noted above> an earlier use of stone is rendered possible in view 
of Rgveclic references to stone-built bulwarks and iorts (vide 
ante). The very word 'Smasana* (possibly from ' asina-sayana/ 
according to Weber) would suggest that stone was all along the 
chief material in its construction ; so that the origin of this 
special type of funeral structure would be Magadhan and non- 
brahmanical, and when other materials are used, this would be 
due to brahmanical adaptation of the ' smasana/ characterized 
by opposition to use of stone and adherence to their OWB 
traditional bricks (vide ante). This view would also agree with 
the fact that the Sat. Bra. does not give details of the * grhan/ 
* prajnanam * and ' round ' forms of the ' Smasanas,* and that 
whereas the symbolism of the altar is specially brahmanical, that 
of the s smasana ' is Buddhistic. 

2 Made of * vrtra '= stone, ace. to comm; the timber pillars are made 

of ' palasa/ * sami * and * varana ' : Sat. Bra. XIII, 8. 4, 1; (of- 
Kat. gr. But. XXI, 3, 31, and Sat. Bra. IV, 2, 5, 15, with 
S-B.E 44, 437, n- 1). * Sanku ' being associated with a tapering 
form, the stone-* sanku * would have a gradually narrowing shaft 
(like an obelisk). 

3 Of. the 4 pillars adjacent to the * s'fcupa/ an,d later on to medieval 


4 Roadside pillars and gateways are referred to very much earlier in 

the Av. (XIV, 1, 63) where marriage processions pass along the 
well-made road through 2 pillars (asked not to injure the bride: 
hence hitrh and heavy) ; an arch or ' torana * is evidently implied; 
these pillars (sthuria) may well have been of bricks or stone, 
Cf. Av. XIV, 2, 6, *" pillar standing in the way," which however 
might refer to a row of posts barring the road, 

5 Vide n- 3, &. 34. 

6 Cf. the roya. name 4 Drupada/ beside ' Banda,* * Danda-dhara/ etc., 

* found in Puranic and Epic lists; cf, also the ancie-nt name 

7 Cf. the symbolism of ' danda ' and ' skambha/ 

8 The keynote of Buddhistic (and Saiva) architecture is this * smasana * 

symbolism, just as the ' altar ' symbolism? is associated with 
brahmanical structures; it is probable that In Taitt. Sam. V, 
2. 8, 5* (p. 42, n. 2), these two ancient groups of symbolism 'are- 
hinted at. ,.... 

9 As'okan inscriptions refer to widely distributed pre-existing monolithic 

pillars in which he ordered Ms edicts to be inscribed (cf. 
Samudragupta) ; vide end of Min. B. Ed. I, Bup. Tea*; end of 
Pill. Ed. VII; as opp. to fresh erection of snch pillars, e.g., 
Rummin. Pill, Inscr t 

( 46 ) 

sufficient explanation of the Asokan pillars, and a theory of 
their Persepolitan origin is unnecessary. 1 

From all this it may reasonably be concluded that a stone 
structural style with round forms, the immediate source of the 
Buddhistic "architecture, was early developed in non- 
brahmanical areas, particularly in Magadha; traces of which 
may be discovered in the earlier Vedic literature (dr. 10th 
cent. B.C. at least), and which was definitely flourishing in 
the 7th cent. B.C. 

Summing up the evidence on structural forms, it seems 
probable that there were three main sources from which the 
early and later Vedic styles, the prototypes of subsequent 
well-known ones, were derived : the Lower Gangetic regions 
(including the delta), the Deccan borderland (including 
Magadta), and the Middle Himalayas (with submontane 
areas) . These regions quite naturally gave rise to building 
styles characterized by bamboo and brick, stone, and timber, 
respectively. The first is associated with Angirasas, 
brahmanism, and what may be called Manva regions ; the 
second with ^ the Vratyas and Magadhas (Pracya ( s) , occupying 
an area assigned by tradition to a stock different from the 
Manvas and Ailas but with superimposed layers of Alias ; the 
third would be brought by the Ailas into the plains from the 
Mongoloid mountainous areas they passed through and came 
in contact with. The ethnic and historical significance of such 
indications in the Vedic literature cannot be over-estimated, 
being also in agreement with the facts of Puranic tradition,* 

1 It seems probable that Mauryan monolithic pillars had their origin 
from the indigenous toddy-palm. Magadia is thickly set with 
palm-groves, the prehistoric prototypes of ancient village halls 
with palm posts and of the Mauryan 1000-pillared halls (at first 
of timber). The palm leaf is of course the prehistoric material 
for writing in the Gangetic valley or the littoral ; and the regular 
lines and spaces on the stem of the palm tree afford ready surface 
for inscriptions or public and royal orders in writing (at first 
with paints) this being suggested by the common use of palm 
leaf for writing. The palm develops a tapering monolith-like 
stem, crowned by a tuft of fans (some branches being often 
cut away for toddy) (Hair-drawing), resembling lions j 'manes 
at twilight and thus suggesting a four-faced lion-capital, while 

iU" e t if > to 5 d /; ves l> el v h ^ g U P aloft would &&^ th s * 

called bell -capital. Probably criminals were hanged on or 

^li M ??!?", tr ? es ty ro ? al order (<* the Vedic and Epic 

drapada ) ; taji is again ' varum,' belonging to Varuna, the 

god of justice, chastisement and kingly power f and the "toddy. 

drawer is as much a Pasi as Varona himself 5 his caste being 

?Mf /T^V 86 ^ peculiar 'P a * a ' (o f Palm-fibre, witH 
the help of which he climbs up the tall slender trees) - with 8 uch 

* y S ? 0]h \ < ?> doubtless, the criminals of old 
T han ^d from the palm trees (a folk tradition 
to be responsible for various apparently unmeaning 

7 ' 

and Si B! Jl J ^ th . at .^ osts and spirits dwell on 
v-/ ! P ?T^ rSo ? s v^^rmi? to rest under them), 
Vide Pargiter AIHT, chaps. XXIV, XXV. and XXYI 


If references to house-building in the Vedic literature are 
few and fragmentary, those to the internal equipments of .such 
structures are necessarily ,so, The details found in the tests 
are mostly connected with ritual, and it is only incidentally 
that some secular and ordinary feature of house-furnishing is 
noted. The^ ritualistic types of furniture, again, cannot be 
taken as a faithful counterpart of the contemporary secular 
ones, for, it is well-known that sacrificial and ritual requisites 
almost always remain primitive and unchanged throughout 
long ages, and it is particularly true of India 1 ; so that the 
4 furniture/ of the priestly texts is almost that with which the 
' Jbrahmanic * cult and civilization started. 2 So great is the 
ritual conservatism in these respects, that even where special 
circumstances required alteration in the sacrificial parapher- 
nalia, the external items are transformed into * brahmanical '- 
looking accessories, by the employment of primitive materials 
sacred in ritual tradition. 3 

Naturally the ' furniture ' most alluded to consists of 
various seats and beds. These were of very different grades of 
comfort and structural complexity, items connected with the 
ritual Jbeing always much cruder. Thus, ' prastara/ 4 a 

sacrificial seat, consists only of strewn grass (darbha) ; 
' barhis/ 5 for the ' seats of the gods/ is a litter of * balbaja ' 6 
grass strewn on the sacrificial ground ; ' kurca 3? is a bundle 
of reedy grass for a seat, or a small square grass-mat easily 
rolled into a bundle; even where a 'cushion-seat' (' brsi% 
' vrsi/ or ' vrsl *) 8 is used, it is of grass. 9 

1 Tha same materials and shapes being mostly retained. 

2 The materials employed would indicate that a good part of the 

Bra-hmanic equipment was * Gangetic/ evidently forming 
the original stockj which was supplemented by other acquisitions 
of a Himalayan and middle- country character. (It would seem 
as if the ritualism of the Brahmana age had inherited the 
traditional * equipments ' of both the Northern Aryans and the 
Eastern Pre- Aryans). 

3 E.g. the ' Brahmana ' treatment of the Imperial throne, sadly 

reduced and metamorphosed in the ritual. 

4 Rv. X, 14, 4; Av. 2, 6; Taitt, Sam. I, 7, 7, 4; Vaj. Sam. II, 18; 

XVin, 63; Ait. Bra. I, 26; H, 3; gat. Bra. I, 3, 3, 5; etc. 

5 Quite common in Kv,,, Taitt. Sam., Vaja. Sam, etc. ; (vide Y.I. 

II, 61). 

6 Kath. Sam. X, 10; Taitt. Sam. II, 2, 8, 2; Mait. Sam. II, 2, 5. 

7 Taitt. Sam. VXI, 5, 8. 5; Sat Bra. XI, 5, 3, 4.7; Ait. 'Iran. V, 

1, 4; "Brhad. Up. II, 11 ? 1. 
S Ait. Aran. 1, 2, 4; V, 1/3; 3. 2; Sankh, 8r. Sut. XVII, 4, 7; 6, 6; 

Kat. gr. Sut. XIII, 3, 1. The_ * brsi ' seat, i.e. the padding of 

it was & span high (Sankh- Aran. (Keith), viii). 
$ Just as the sacrificed wife wears a garment of Kua grass for 

rites, a relic of primitive dress (Sat, Bra. V, 2, 1, 8). 

( 48 ) 

But there were other seats of a more advanced type, 
Thus the ' sadas,' from which the * sadasya J1 watched the 
performance of the sacrifice, must have been a raised seat, 
and of a style specially associated with his office. Ihe 
'kaffipu' 2 is a mat or cushion made from reeds ( na4* ) 
crushed by stones, and ' nadvala ' 5 is a bed of similar stuff ; 
and * kata ' is a ' vaitasa * or rattan mat, made of split cane 
or cane-like bamboo. These were the products ot regular 
ancient and indigenous crafts : ' kasa ' 5 was very early used 
for mats, etc. ; and there were professional women workers in 
1 nada ^ (reeds, canes, etc.) of the swamps, or in * kantaki, 
apparently the thorny cactus, whose fibrcs were used to plait 
mats and stuff cushions. 8 These ' karl 's evidently turned out 
artistic seats and carpets, as the early occurrence of ' hiranya- 
kaslpu IS shows ; the reference here is plainly to the use of 
4 gold threads and fringes ' in the web, borders and designs 
of the ' mat ' ; so also the ' golden kurca no on which the King 
sits at the * ASvamedha,' while the ' thotr ' sits on another 
4 golden ' seat ( * kafipu ' ) , Is clearly the finished rich work - 
of craftsmen, a-s compared with the primitive bundle of 
plaited grass. 

It Is noteworthy, however, that all 11 the ' seats ' mentioned 
in the ritualistic tests, are made of long grass, reeds or other 

1 Tide n. in S.B.E. 43, 348, re Sat. Bra. X, 4, 2, 19 (cf . Keith. 

Ait. Aran. 37). 

2 Av. VI, 138, 5. It Is noteworthy that in Tamil ' kacci '= reeds or 

stalks, and * pa ' means mat so that ' kacci-pa ' represents the 
original of this 4 kaslpu ' ; (cf. also * kacci ' = creeper and 
eocGanut-shell fibres, and ' kaccu *= fibre or grass-ropes, with 
which cf. Tern, 'kachi'j. 

3 Vi,ja. Sam, XXX, 16- Taitt. Bra. HI, 4 } 12, 1. 

4 Taitt. Sam. V s 3, 12, 2; cf. gat. Bra. XHI, 3, 1, 3. 
& Bv. X, 100, 10 j Taitt. Iran. VI, 9, 1. 

6 Av. VI, 138, 5 j * na^a ' growing in lakes and in rainy season ; BY. 

VIII, 1, 33; Av. IV, 19, 1; ('na<ja,' is frequent in Av., Yv-j 
Bra. and Aran.). Cf, note 7. 

7 Vaj. Sam. XXX, 8; Taitt. Bra. HI, 4, 5, 1. 

8 * Thorns'* (vide V.L, I, 133) could scarcely have been made into 

cushions and nsed to plait mats. (The nse of cactus fibres for 
weaving or similar purposes is indigenous in many isolated 
districts, even to-day). 

i Av. V, 7, 10 (as an adj., used of ' Arati * conceived of as a gorgeous 
woman or a courtesan, with golden mantle (drapi), etc.) ; also in 
Taitt. Bri. Ill, 9, 20, 1; Ait. Bra. VII, 18, 12; gat. Bra. XIH, 
4, 3 f 1. (It is not necessary to suppose another * cloth of gold * 
spread over the * kasipu,' which itself could be * golden * in the 
above sense. 

10 Sat. Bri. XHE, 4, 3, 1 ; ace. to the comm, it was a golden stool 
with feet, having a * kurca '-like pad over it (* pltham 
karcSkfti/ which might also mean * a wooden seat, carved or 
painted* ""by * alpsna,' so as to resemble a * kurca,' i,e., in view 
of suitability to ritual). * 

U Except probably the * sadas * of the * sadasya * (vide ante), which 
may have been a raised seat of some sacred wood; but the 
occurrence is not very early ? and the 17th priest was rathe? 

( 49 ) 

products of riparian lowlands, where alone the industries 
alluded to could hawe flourished 1 in that early age. A 
characteristically Gangetic outfit would thus seem to have been 
the stock with which the ' brahman ' ritualism started. 

But 'beds,' 'couches for reclining,' and other 'seats/ 
which had little connection with the sacrificial ritual, are of 
woodwork principally. Thus the ' pitha ' (alluded to In the 
mention of the ' pithasarpin ' 2 cripple) was ^evidently a wooden 
seat; and like its later representative * pi<Ji ( ' T pitjUha/ etc.}, 5 
it must have been a low, rectangular, polished seat (some- 
times carved, and oftener painted with designs). The 
' talpa ' is made of * udumbara * wood 4 (heavy and strong) , 5 
with four feet and four frame-pieces ( ' usyala ' ) ' fashioned 
by Tvastar * (i.e. carved and moulded by skilled carpenters), 
and with embroidered, and inlaid (' pi ') * vardhras ' (straps 
of leather, etc.) in the middle of it. 6 The * prostha s ie clearly 
wooden, 7 and the * vahya ' at least partly ( so. B So also the 
* asandi/ which the Vratya chief uses, is a comfortable chair 
ef wooden framework 9 with adjuncts of diverse other 
materials ; and the * asandi ' for the King in the ritual is 

is not directly connected with ritual till the time of the Siitr&s* 

though they ?re known much earlier (vide infra, and n, 10, p. 48; 
(an * udumbara * stool is u&ed in Sankh. Aran : (Keith i x) by 
the Udgatr in the Mahavrata). 

1 As they flourish at the present day, in the Gangetic districts of 

Bihar and Bengal, where these crafts have almost become arts, 
with an ancient tradition. 

2 Vaja. Sam. XXX, 21 ; Taitt. Bra. HI, 4, 17, 1. 

3 This is a characteristically Eastern furniture, and the linear designs 

painted on it (the famous * alpana ') are also of Eastern origin 
and _development. (The word * pitha ? may be a Sanskritisd 
pre- Aryan one). (The ' alpana * would explain how the surface 
of the King's! golden ' pitha * (vide n. 10, p. 48) could be made 
to look like a seat of * kurca '), 

4 Taitt. ^ Bra. I, 2, 6, 5. 

5 That is the reason, the Brahmanas give for its employment in the 

King's seat * asandi : but it is more of a reason to connect the 
use of f udumbara ' for the secular- * asandi ' and * talpa * with 
the fact that this wood is indigenous to the sub-Himalayan 
tracts. So also, other * asandl's (e.g. of the Bharatas) are made 
of * khadira ' wood, also indigenous to the same region ; with 
this is to be compared what has been said above re a Middle- 
Himalayan * timber ' style 'as opposed to the primarily 
* brahmanic * * bamboo * style. 

6 Av. XIV, 1, 60 : the descr. is understood of the * <talpa * by Kaus- 

76, 25, probably correctly; ifc does not apply very well to the 
bridal ' car * in the next verse (though usually taken in that 

1 Cf. the descriptive epithet * prostha-pada ' {a name in Jaira* Upan. 
Bra.). 'Cf. also the corr. vern. form ' paitha,' 'a wooden bench, 
a broad plank resting on two legs or two vertical planks, 
specially used of the rowers' benches in the Gangetic river-boats, 

*' Vide infra. 

9 E.T. in Av. XV, 3, 2ff, 

( 50 ) 

similar*' but hare the woodwork, etc-, seem to have been 
replaced with or supplemented by grass, reed, * 
Sne-work/ in conformity with ritual tradition.* 

The ' beds * or ' couches y mentioned, all belong to the 
equipment of the innet apartments of a house, being connected 
with women, ' Talp& '* is apparently the nuptial bed- 
stead, 5 used by married pairs only, as the special us^ot fche 
word in ' talpya * (' legitimate son/ being born in the 
nuptial bed) and * guru-taipa/ 7 and its being made of the 
sacred ' udumbara/ indicates. Some women in a big house 
(' harmya ') are described as ' prospia-saya/ 8 reclining on a 
1 prostha/ where something like a high and broad bench may 
be meant, as elsewhere, 9 being distinguished from * talpa 
and ' yahya ' ; apparently it had strong moulded and turned 
legs, for ' ; * prostha-pada"* was a proper name. 10 It seems 
probable that such long timber seats were fixed against the 
walls, 11 or were combinations of a settee and a coffer ; thus a 
coffer (' kosa ') with a pillow 12 is sent along with the bride when 
she goes to her husband's home': .such marriage-coffers 
evidently could be used as couch or bed 13 ; and it is noteworthy 
that both these types of ' bed ' are found in the inner apart- 
ments of the middle-Himalayan villager's dwelling-house. 14 

1 E.g. in Ait. Bra. YIH, 5; 6; 12. 

2 E,g. in Sat, Bra. XII, 8, 3, 4ff; XIV, 1, 3, 8ff; VI, 7, 1, 12fL 

5 ** Because reed-grass is meet for sacrifice *': Sat, Bra. XII, 8> 3, 4-10j 
XIV S 1, 3. 83; the process of ( brahmanization ' is clearly 
indicated by the direction in Sat. Bra. HI, 3, 4, 31, wher 
all the * human * particulars of the * asaadi ' are forbidden ta 

be imitated in the ( ritual J * asandi.' 

4 Ev. VU,-55 3 8=Av. IV, 5, 3 (vide infra, for sense of * talpa ' here) ; 

Av. V, 7, 12 (king and his wife's) ; XIV, 2, 31.41 (bridal) ; 
Taitfc, Sam. VI, 2, 6, 4; Taitt. Bra. II, 2, 5, 3 5 Pane. BrS. 

XXXH 3 4/2; XXV, 1, 10. 

5 Corresponding to the * viyer khat * of Bengal, to which a peculiar 

sanctity and significance is attached, and which may only be 
used by the married pair who first used it. 

Sat, Bra. XHI 3 l } 6, 2. 

7 Chand Upan. V, 10, & 

8 Kv. VH, 55, 8=Av. IV, 5 a 3. 

9 Taitt Bra. H, 7, 17, 1. 

10 Of. a. 7, p. 49- 

11 So that the * prostha ' having two * padas * only (cf, n. ?, p. 49) 

would afford a parallel for men's legs. 

12 AT, XIV, 1, 6. 

13 Cf. the medieval Germanic marriage-coffers (of woodwork), which 

re very mnch like the combined bed and coffer of tb* 

Himalayan houses. * 

14 E.g. in the timber-bnilt houses of the Simla Hill States, where 

these are used by women-folk for naps between work s or as 

regular bedi. 

si y 

' Vafaya n is <a conch of a comfortable kind, osed by women ; 
the name suggests a light structure, that could be carried 
about when necessary, so that it would seem to have been a 
canopied reclining arm-chair, with poles or handles for 
carriers. 2 But a ' vahya,* * bearing all forms J (i.e. of carved 
wood-work) , and with a gold-embroidered coverlet ( rukma- 
prastarana * ) , is the bed on which the bride mounts and lies 
with her groom in the marriage-ritual. 3 This seems to be 
referred to in the next ' mantra J as th-e 4 talpa f of the pair 4 ; 
and after the consummation the ' demons * of this * talpa * 5 
are got rid of by the priest. Thus the- bridal * vahya J would 
be something more than a mere litter or sedan-chair, a 
regular bedstead, capacious enough for two ; so that * vahya * 
might be taken to signify the bed carried along with the 
bride 6 to her new home as part of her dowry. 7 But this again 
is rendered uncertain by a following consummation- 
6 mantra/ which shows that during the ceremony the couple 
had also lain together on an ' asandl ' 8 (settee), with cushion 
and coverlet 9 ; this * asandl * cannot have been a full bed. 10 So 
the bridal ' vahya ' need not be taken as identical with the 
* talpa * mentioned in the same connection ; and it would 
rather appear that the * vabya, 1 ' talpa/ and ' asandl/ were 

1 -Hv. VII, 55, 8=Av. IV, 5, 3 (women sleeping on it); Av. IV, 20, 3 

(weary bride mounting it) \ XIV, 2, 30 (used in marriage 

2 Something like the modern ( dandi * of 'the lower Himalayas or 

* duli * of the plains, also used by women mostlv. 

3 Av. XIV, 2, 30. 
* Av, XIV, 2, 31. 

5 Av. XIV,, 2, 41. 

6 It would of course be something distinct from the * c< ko3a* and 

pillow " similarly sent with her {see above). 

7 Cf . the same custom nowadays. (For ' vah ' in the sense of * bring- 

ing dowry, etc., along with an>e, s cf. a King's wife called 

* Satavahi J : Av. V> 17, 11. 

8 Av. XW, 2, 65. 

a Or the * upadhana * and * upavasana * might refer to the already 
used * talpa ' and covered ' vahya * respectively (vide ante\ ; 

* upavasana * might also refer to the dress of the bride herself. 
In any case, the * vahya/ * talpa * and * isaadi ' are all need 
by the couple. 

10 Elsewhere in Av. and in Ait. 'and Sat. Bras., 'Ssandi* is a throne 
or throne-like seat (vide infra) ; but once in Sat. Bra. (vide 
infra) and in Buddhist tests (cf. Digha Nikaya, H, 23) it is 
said to be carried by 4 men (implying a longish reclining chair) j 
and ' asandia * in Hala is later on glossed by * parvankika * and 
*khatva,' * pointing to a long couch; but in earlier literature 
' asandl ' is definitely a * seat/ an^ & * IS rather the * vahya - 
^ which corresponds to a * long reclining couch.' _ 

* (Quoted in Whitney and Laninan, Trans, Av.) ; the paryankika 
is comparable to the * prati^ayyika v of Vats. Ki, Sat.; 
* JchatvS * is a light, narrow, cord-or strap- woven 

( 52 ) 

three essential items of furniture for the bridal chamber. 1 So 
also, the ' vahya ' is specially associated with the bride, as 
shown by the incidental simile ; ' ' like a tired bride ascending 
the ' vahya,' " 2 evidently referring to the above marriage 
ritual. Thus the apparently obscure distinction made In Bv. 
VII, 55, 8, become clear : it refers to married women occupy- 
ing their commodious * talpas/ the new bride (or prospective 
brides, one of whom is sought to be approached secretly) on 
the fashionable 4 vahya,' and other single* women of the house- 
hold on the sterner 'prosthas/ within the * harmya * or big 
family-home. / Sayana ' 3 is a general term for bed or couch, 
with no particular features, except softness and association 
with women, 

A number of details are given about the * asandl * (and 
'the ' paryanka ' ) : apparently because furniture of this typo 
was not common in the ordinary priest's dwellmg-houHe, and 
originated with the ruling nobility, 4 though in their ritualized 
and modified form (reed-covered and clay-daubed), 5 these 
must have been subsequently used by brahmaiis also. 6 

' Asandi/ literally, 7 is a generic term for seat of .some 

1 Another item would be the * kosa ' and pillow brought by the 

bride; the red ox-hide spread over strown * balbaja ' (rush), 
on which the bride sits, is part of the ritual requisites (Av. 
XIV, 2, 22-24). 

2 Av. IV, 20, 3. 

3 Av. Ill, 25, 1 (of a beloved woman, a, maiden ; * ut-tuda * is an 

unexplained word in this verso ; can it mean * silk * or * silken 
coverlet* of the maiden's * sayana,' from * tuda 8 = mulberry 
leaves (i.e., sprung from ' tuda')? The meaning would then be, 
"let the silken coverlet on thy bed, pain (*tud') thee," etc., 
' involving a pun on tuda '). Also Sat Bra. XI, 6, 1, 2 
(Pururavas and UrvasTs couch) ; ibid. 7, 4 (aof t couch of a 
Vedic student); Av. V, 29, 8. 

4 Thus the^ asandi is called the ' navel ' and 'womb ' of * rajanyas,* 

and is always specially characteristic of .the ruling chief. 
(Even to-day * palang * (or *palanka') is more aristocratic than 
'khat* (' khatva '). 

5 Vide infra. 

6 Thus, such a, seat 3 with cushion, is prescribed as ' fee ' after funeral 

rites : Sat. Bra. XIII, 8, 4, 10. 

t * Asandl ' should properly mean either a ' brilliant seated person * 

or .a * shining seat/ i.e., a throne as well aa an enthroned prince 

(this sense is perhaps al&o implied ia the name of the Kuru 

capital * Jisandi-vant * ; vide ante). ( Di * in ' asandl * is indicative 

of lustre or prominence ; or perhaps ' andi ' may be an early 

Sanskiritic suffix indicative of prominent and ever-present features ; 

cf. words similarly formed: * vasandi * (domiciled), corrupted 

into colloq. vern. ' vasunde * ; ' bhusandi * (gaudily, nncouthly 

dressed), corr, into colloq. vern. ' bhusuncji ' ; ' kalandi * 

(rippling), changed into * kalindi, 7 a river name! It seems likely 

that AsandMmitTa, q. of A^oka, was so designated being * mitra * 

or consort on the * asandi * or throne,' Asandi-initra * being the 

regal title. 

'( 53 } 

fine sort, but from its first mention onwards, a special type 1 

of seat is almost always implied by it ; the type varies in 
different references, but the earlier >and more usual form 2 is 
something like a comfortable ' gadi ' (equivalent to a throne), 
that) might be used by the ruling aristocracy or on special 
occasions by other people; a secondary and modified form 3 is 
that adopted in ritual, where a king is concerned, or where a 
deity is conceived of as a king ; still later is the form more or 
less approaching a bed, referred to in Pali and early Prakrt 
literature. 4 

The * asandi ! is first referred to in Av., and in connection 
with the inauguration of 4 the Vratya } for whom it is brought 
together 5 ; and as the origin of royalty is there ascribed to this 
idealized event, 6 it w r ould appear that the Atharvanic tradition 
regarded the first kings as ' Vratyas * (in all likelihood 
Easterners) 7 and the l asandi ' as the royal seat specially 
associated with them. 7 It is to be noted here that 4 asandi ' 
does not occur in Ev. 3 8 though allusions to things * regal ' are 
not altogether wanting in it ; the* force of this point, however, is 
weakened by the references in the Brahmanas to an ' asandi ' 
" like those of the Bfaaratas," 9 a * Bgvedic ' midland dynasty, 
and to an older * gatha * mentioning c Asandi vant/ the capital 
of Janamejaya-Pariksita I of the same race. 10 Thus the 

1 This comes out clearly from the comparative summary of descriptions 

given infra. 

2 From Av. onwards. 

3 Particularly in the Brahmanas. 

4 Vide n. 10, p. 51. (The ' asandi ' probably \vas displaced by 

the subsequent ' paryanka ' (vide infra) and the c sirjihasana ' ; it 
is not referred to after the 1st century A.D. ; it is poasib.e that 
the latter correctly represents the earlier * asandi ' which was a 
' vyaghrasana * (with tiger-skin spread neck in front) : i.e., the 
Eastern and Gangetic style of * throne * was modified by contact 
with West Indian conditions, the ' tiger * symbolism being re- 
placed by a ' lion ' one j vide next para, and notes 9 and 10 

5 Av. XV, 3, 2ff. . , i_. 

6 The emphatic view of 'she Av. about the e vratya ' origin of kingship 

and priesthood, and the great political power and prestige of 
the ' vratya,' finds complete support from the ^ Puranic tradition 
regarding* -them, and seems to be only a priestly and mystic 
version of accepted and known facts of that tradition : the 
1 vratyas * corresponding to -the non-Ailas. Tha ' vratya * hymns 
can be much better explained by this reasonable hypothesis, 
than by supposing that the ' vratya * is a wandering * sadhu ' 
or a pretentious ' saiva > mendicant, or a personification of 


7 See n. 6 above- 

8 But * upavarhana ' (and nl) and upastarana are known to Rv., 

and these were particularly connected with asandi ; cf. Kv. ^ 

85, 7; IX, 69, 5. " , . . . , 

9 Sat Bra V. , 4, Ifi. The Bharatas were however much influenced 

'by the Ingirasas, ace. to Pur. tradition. ^ 

10 Ait. Bra. VIII, 21. This king is placed by Purinic tradition about 
20 steps before the close of the Rgvedic period. 

( 54 ) 

* trltya ' (Eastern) emblem of royalty (as known to the 
Afidrasas) would appear to have early been adopted by 
Midland rulers also. The same original .connection with 
Atharvanic tradition is probably indicated by the use ot the 
' isandi ' in the Av. marriage-ceremonial, 1 and its absence in 
that of the Kv. 

The ' paryanka '' is a later development, being first 
mentioned and" described in the later Yedic texts 2 ; it is a 
magnified * asandi,' and like it associated with regal style and 
opulence, rather approaching a bedstead, but yet used for 
sitting only; so also, later on, ' asandia ' is taken to mean 
iSa/ a smaller * paryanka.' 3 

The general type of these ' high class ' seats "comes out 
sufficiently clearly "from a comparative summary of descrip- 
tions in the texts: (i) In Av. : (a) 4 The Vratya chief's 
4 asandi ' : framework of wood and cording or straps ; 2 (fore) 
feet, 2 (back) feet 5 ; 2 lengthwise and 2 crosswise pieces ; for- 
ward and cross ' tantu's (rather ' woven ' straps, than 
* cords ') ; and ' upasraya/ the support or back of the seat ; 
adjuncts: e astararia/ coverlet; * asada,' seat proper (i.e. the 
cushion for sitting on) ; and * upavarhaija,' cushion for leaning 
against. (b) & The bridal '-asandi': the framework is not 
described; it may have been a bed-like reclining couch, 7 but 
the supposition is not essential 8 ; adjuncts: not clearly 
defined ; the 4 upadhana ' (pillow) , and * upavasana * (cover- 
ing cloth) may or may not belong to it. 9 

(ii) In Yv. Samhitas 10 : though often mentioned, 
descriptions are rare 11 ; here also, the * asandi * is specially 

associated with kingship or imperial rank, 12 and -secondarily 

j 2, 65. Tlie Rgvedlc marriage ceremonial (in. its last book) 
is only a * selection 1 from the Atharvavedic one, which must 
be very much older and traditional. 

2 Kaiis. Upan, I, 5. Of. Jaim. Bra. II, 24; and Saukli, Aran. iii. 

3 Tide a. 10, p. 51. 

4 Av. Y S 3, 2 ft 

5 This distingmsMng of feet probably points to a rectangular frarn. 

Av. XIV, 2 f 65, 

? So as to suit the marriage ritual better, 

8 Thus, ^tfee use of capacious * throne-seats * for the newly married 
pair ia Ia!iaa ceremonial is traditional. 

a Vide &. 9> p. 51, 

10 Taitt, Sam, VH, 5, 8, 5; Vaja. Sam. VIII, 56; XIX, 16; 86; etc. 

11 Skin coTer and smooth and pleasant seat : Taittr Sam, I, 8, 16. 

12 E.g. Vaja, Sam XIX, 86^ where ths ' asaudi ' is regarded as a 

"mother/ f.. the * womb of raianyas * (aa elsewhere, e.g. Vaj* 
Sazp. XX, 1). 

{ 5 ) 

with gods, 1 while its use in ritual by a sacrificing priest 2 
ensures * samrajya * for his client ; but elsewhere the 
qualificatory term * rajasandl ' 3 shows that humbler 
4 asandf s ' were in use amongst other people at the same 

(lii) In the Bralimams: (a) In the Aitareya : the King's 
1 asandi J adapted for use at consecration and other * regal ' 
ceremonials ; (a) 4 Quite a small seat (evidently for temporary 
use during ritual) ; framework of ' udiunbara * "wood ; the feet 
a span high; the * head-* and cross-pieces each a cubit (i.e., 
a 'square* type) 5 ; the interwoven part (* vivayana ') of 
plaited ' munja ' reed; adjunct; ' astarana/ spread, being 
a tiger-skin, placed neck in front (so that the long skin would 
cover both the (sirsanya * and the seat proper). (f>) 6 Frame- 
work the same (of * udumbara ' and with * sirsaiiya *)j but 
the specification of front feet and back feet .shows a l rect- 
angular ' type (with probably differently moulded pairs of 
legs) ; and the lengthwise cords and cross-ties are apparently 
run through holes 7 in the frame-pieces; adjunct : 4 upavarhana/ 
back cushion, (c) 8 Another description: same framework and 
other details, as in (b). (b) In the Satapatha : (&) 9 * Asa-ndi ' 
4 'like that of the Bharatas,'* and specially a 'rajanya ' 
seat (being the * womb ' of that class) : a high seat 
above the level of low seats of surrounding subjects ; 
made of * khadira ' wood, perforated (' vitp^a), and 

1 E.g. Vaja. Sam VIII, 56, the seat of Varuna (conceived of usually 
as a great King). 

3 Taitt. Sain-. VII, 5, 8, 5. Two other seats are used at the sama 

time,, the ' kurca ' and the * plenkha,' winch last can hardly have 
been, an ordinary i swing/ (In the Mahavrata ceremonial (as 
in Sankh. Aran.), the * swing ' is set up on timber-posts no 
doubtj but is used only as a ' seat * ) , The comm. gives * dola * 
as its maning ; in vern. * dola * is the same as * $ u ii>' a sort- of 
carrying chair, which doos swing ; * dola ' and * dull ' are used 
indifferently in Bengali. The sense of a rocking chair * is 
however admissible. The comm* here glosses * asandi * by 
* khatvakara/ 4-legged, and high. 
5 Vaja. Sam. XIX, 16. 

4 Ait. Bra. VIII, 5 and 6. (The seat is mounted with the _ right knee 

first 3 then the left, approaching from behind and taking hold ol 
it by both hands). 

5 Of. the modern * khatli ' or * carpal y of Upper India, characterized 

by the same span and cubit measurements and square type, with 
4 moulded and painted legs, and the ' seat ' of stretched woven 

S Ait. Bra. VIII, 12. 

? This shows acquaintance with cane-woven seats ; thus there w$re two 
main types of * seats ' in these * asandls,* with cane {or fqui- 
valent) run and woven through holes in frames, or broad straps 
(leathern* or woven stuff) wound over and across the frames. 

S Ait. Bra. VIII, 17. 

tSat- Bra. V, 4, 4, Iff, 

( 56 ) 

joined with straps (' vardhra '), pleasant and soft- 
seated, and placed on a tiger-skin. (6) 1 The ' imperial ' 
and * ksatra ' ' asandi ' adapted for sacrificial ritual : 
made of ' udumbara ' wood ; knee high ; of great width 
and depth; covered with plaited reedwork, because reed-grass 
is meet for sacrifice, -and for the same reason, the ' spread ' 
irs a black antelope-skin. 2 (c) 3 ' Asanil ' of the c sainraj/ 
similarly adapted: of * udumbara/ and shoulder-high (as com- 
pared with the raja's navel-high ' asandi ') 4 ; wound all over 
with cords of rush ('balbaja') owing to ritual mystic 
significance, (dp ' Asandi s used in pure sacrificial rites : of 

1 udumbara ' ; a span high; a cubit in width and depth (i.e. 
of a 'square ' type) ; covered with reed-grass cords, and 
daubed with clay as well. (e) 6 ' Asandi/ said to be also 
called ' rta-sadani ' (throne of justice) , ascribed to a deity in 
ritual : of ' udumbara ' ; navel-high ; to be taken up by 4 men, 
instead of 2 who ordinarily take up the King's * asandi/ 
many details of which are explicitly stated to have been 
dropped or modified, as "human elements are to be eschewed 
as far as possible in sacrifices, >J 

Civ) In Sankh. Iran, and Eaus. Upan. 7 : (a) Brahman's 

far-shining ' c asandi ' (in an extensive fcflfl, of an invincible 

abode, m^ a city) : 2 fore feet, 2 hind feet, 2 lengthwise and 

2 cross pieces. This is evidently regarded as a smaller and 
minor seat beside the * paryanka ' next described. 
lb) Paryanka of unmeasured splendom: ' : same arrange- 
inent of feet and frame, and straps (' tantu ') stretched 

1 Sat. Bra. XII, 8, 3, 440. 

2 Of. Asandi of ud^mbara J with spread of gaat-skin : fa. Bra. 

3 fat. Bra" XIY/1, 3,8 ff. 

4 Sat, Bra. in. 3, 4 3 26 ff 

5 Sat. Bra, VI, 7, 1, 12 ff . 

6 Sat. Bra. Ill, 3, 4 5 26 ft. 

1 Kans. Upan. I, 5; Sankfa. Iran. m s (cf. Jaim. Bra. H 3 24). 


Though the Vedic references to the materials and manners 
of dressing, etc.,, are few, yet incidentally they throw much 
side light on contemporary -social conditions. Thus a quite 
evident feature is a considerable ^ 7 ariety in these materials 
and manners, which can only have developed with different 
regional conditions and tribal customs and tastes : so that any 
general reconstruction ,of one typical Vedic or Indo- Aryan 
dress, etc., from those references, would be more imaginary 
than .scientific. 

Skins form one class of * Vedic * clothing material. The 
Maruts .are dressed in deer-skins, 1 and the gods alarm the 
enemies with coats or shields of such skins (* hari^asya- 
jinena ') 2 ; ' muni's wear brown and tanned skins ( { pisanga 
mala y ) 3 ; and skins of black antelopes are in common and 
traditional ritual use in the Av.. and Yv. 4 It is noteworthy 
that none but ' gods ' and brahmans use skins ; the only 
exceptions being the Vratya chieftains and their followers, who 
have an improved style of wearing twofold (' dvisamMtani ') 
" ajina ',s, one black and one white C krsna-yalaksa *), so as 
to ibrm fur-lined skin-wraps, 5 a-nd the aboriginal forest tribes 
(evidently Ivolarian) who wore * krtti's (and * dursa ') at 
dances, 6 and used ' ajina's. 7 Again, it is the goatskin 
(' ajina ') that is primarily and mainly used (all other skins 
being called 'ajina 1 ), other varieties being the skins of 
the * krsna ' (black antelope) and the ' harina ' and ' eta f 
(spotted deer) ; but no sheep-skins, camel-skins, etc., are 
mentioned as worn or otherwise used, 9 On the other hand, 

1 Ev. I, 166, 10; of the 'eta' or spotted deer, hung from th 


2 Av. V, 21, 7. 

3 Ev. X, 136, 2. Of. the brahman priest going clad in * ajina ' (goat- 

skin) according to ritual custom^ Sat. Bra. II I 3 9, 1, 12. (tfe 
tanning. in Bv. and later, vide V.I., I, 257; re the furrier's 
trade: cf. Yija. Sam. XXX, 15; Tartt. Bra. Ill, 2, 13, 1; 
*the skins worn must have been properly dressed). 

4 Vide V.L, I, 185; and of goatskins: cf. Sat. Bra. HI, 9, I, 12; 

V, 2, 1, 21.24 (ajarsabhasva ajinam). 

5 Pane. Bra. XVII, 1-15; cf. Kat. Sr. Silt. XII, lj XXII, 4; Lat. 

Sr. Sut. VIII, 6j Apast. Sr. gut. XXH, 5, 4-14. 
i Av. VIII, 6, 11. 

7 Av, IV, 7, 6. 

8 For other purposes the skins of boars (and antelopes) were used tor 

shoes, and of tigers for seat-spreads; rhinoceros hides for 
chariots; red cow-hides for ritual scats and war-drums. 

9 gat. Bra. Ill, 1, 2, 13 ff. may point to a tradition * wea f T C0 ^' 

hides in primitive ages; e avika* in Kat. Sr. but. AJLI1, <*, 
seems to mean sheep-skin, but it is evidently a late addition of 
Sutrd period, not being found in the corresponding older passage 
in Pane. Bra. XVII, 14-16. 

( 58 ) 

the texts know of the primitive riparian clotting material of 
grass : the * kusa J skirt (round hips only) which the 
sacrifice! 's wife has to wear over her ordinary dress, 1 is 
evidently a relic of the prehistoric grass garment ; with this 
may ba compared the common ritual _ use of grass girdles, 
* sani ' or * maufiji,* first invented by Angirasas^ These facts 
can only signify that -this specially brahinanical and ritual, 
Vratya and aboriginal, use of certain varieties of skins and 
grass-reeds as clothing, arose in the Eastern Gangetic country 
amongst early indigenous "peoples, 3 and is_ not part pf any 
extra-Indian North- Western outfit that Aryan immigrants 
from Central Asia might be supposed to have brought with 

Another material for clothing was wool (urna) . The late 
occurrence of * avika/* sheep's wool, shows* 'that the first 
source of wool in Vedic India was the goat, just as the first 
skins worn w:--re goat-skins 5 ; 'urna/ also, primarily means 
'hairy covering ' of any animal, 6 though the ' urna ' of tihe 
Paras^etc., must refer to sheep's wool. 7 So alsp, there is 
little indication of the divine or traditional use, or ritual sanctity, 
of sheep's wool (or indeed of any wool) : where Pagan is 
called a vaso-vaya ' weaving ' sheep's cloth/ 8 the obvious 
implication is that the ordinary and traditional weaver's product 
was not such ' aheep's cloth ' but cloth of other materials, 
and that it is therefore no reference to the antiquity of 
woollens but rather a glorification of the wool-grower's 
activities beside those of the traditional ' cloth-weaver ' ; where, 

1 gat. Bra. V, 2, 1, 8. 

2 Cf . Sat. Bra, Ill, 2, 1, 10-11 (mekliala > of * sard,' of three cords inter 
_ uvmed with * mu^a/ plaikd like hair, and as soft as ' ur^ia n 

Brhad. Upan. II, 3, 6 ; ' gvika in KSt fe Sut. XXII 4 


'a . 

XVII, m*. corresponding older passage in Pane. Bra 

5 Vide ante. 


s re *-*nfcly iW for 

9 ET. V 52, 9. 

at sacrifices 1 may be simply an unbleached or dyed cotton 
or silken stuff, 2 and not a woollen garment at all. Tie use 
of sheep's wool, then, was not prehistoric, and was almost 
wholly secular, ' as covering (second skin) for men and their 
beasts.' 3 Even so, it does not seem to have been in general 
use. In the Av., ' kambala's (blankets) 4 and ' samulya's 5 
(undergarments of wool ?) are part of the ordinary domestic 
outfit of men and women ; but the * samulya ' may have been 
of ' silk-cotton wool/ 6 and the ' kambala ' of other animal fur 
or hair as well. 7 All the more direct references to sheep- 
farming and woollens pertain to the North- Western corner of 
India only, 8 where evidently it was the staple industry and a 
monopoly. Thus the Indus region was 4 suvasa uraavati/ 

* woolly J and producing fine clothing stuff 9 ; the softest wool 
was of the ewes of the Gandkarans 10 ; Parusiil, also, was 

* woolly ' and produced bleached or dyed woollens 
(' sundhyavah), 11 ' It is to be noted that while Parusnl wool 
Is mentioned in comparatively earlier passages, those mention- 
ing Sindhu and Gandhara wool (further west) are later. 

1 Sat. Bra. V, 3, 5, 21. Mait. Sam. IV 5 4, 3. (pan<Jaraip). 

2 Probably * gairika * or ' gerua,' ol later times, just as the * tarpya 

mentioned in the same connection seems to represent the * tasar * 
sacred to ritual j or it may represent the later ' garad * which 
is pale cream coloured, and goes together with * tasar/ 

3 Vaja. Sam. XIII, 50. 

4 Av. XIV, 2, 66.67. 

5 Bv. X, 85, 29= Av. XIV, 1, 25. 

* Or * salmaliya * : i.e. vests, robes, or wrapper, of light cotton padding 
(like what are ordinarily used even now) ; probably * samulya * 
refers only to a light quilt of cotton-wool, used in- ftlie bridal 
bed (cf . the * kambala ' of 'the bride-wooers). ' (S) amula ' 
occurs elsewhere also: Jaiin. Upan. Bra. I, 38, 4; La$. gr. Sut. 
IX, 4, 7 ; Kans. Sut. LXIX, 3, In the mod. Tamil * seminar! 
= ewe, is probably the original of * salmali' ' to be recognised; cf. 
Tamil ' simbuli'=roDgh cloth, with which cf. the vern. form 
*s(s)imul(a).' 'Vern. ( aamla ? is a kind of narrow shawl (for 
tying round the head or waist); it also means the -embroidered 
end of a turban or * kamarband,' turked or hanging in folds ; the 
word is usually derived from Arabic * shamlat,' from a root=to 
include; but more probably it Is an Urduised form t of the in- 
digenous ' samula,' which would seem to be the original ^of the 
famous ' shawl ' ; cf . the variant Vdic form, ' sabalya/ 

Eg of goats and bears (it is curious that in vern. proverbs -and 
folk-lore the ' kambala ' is made of 4 loma/ hair, and is 
identified with bear-skin. Of- Tamil, ' kamb (p) aji '=mgh hair- 
clfcthj al^o 'simbu^i'; and ' kurunibadu = hairy, fleecy. . 
the Bv form ' simbalam } for * salmali ' : Ev. Ill, 54, 22). 

t Industrial traditions are remarkably persistent in India: even now, 
Ludhiina, Dhariwal, Amritsar, Lahore, Peshawar and Kabul, 
with their typical woollen manufactures, carry on the traditions 
of the Paruni-Gandhara area. 

9 Ev. X, 75, 

11 Bv" IV, 1 22, 2; V, 52, 9; the river was so named from being in a 
wool district. 

( 80 X 

AJ1 these Vedic facts regarding wool become fully 
intelligible When referred to some of the main _facts 
of traditional history l ; this knows of no Aryan 
expansion eastwards from Afghanistan, but, according 
to it, the Ailas (and some Mauvas) a progressed from 
East to West, from the G-angetic country to the Punjab 
and beyond, in gradual and well-marked stages; hence there 
is no indication of an ancient use of sheepskin or sheep's wool 
in the Vedic texts; as the Punjab came to be colonized, a 
specialized wool industry naturally developed ; but there is no 
mention of sheepskins, for the .skin-wearing stage had long 
been left behind, and the traditional vasa-vaya's craft was 
simply transferred from one material to another 3 ; thus, again, 
the Parusni wool came to be known first to interior India, 
and then the Indus and Gandhara products ; the nature of 
the complimentary references in the above passages also 
becomes clear: a Midland rsi aptly apprehends that the 
attractions of the flourishing wool-district of the Parusni may 
have detained his gods ; the high-flown praise of the Indus 
with its wool manufactures (to the exclusion of other rivers and 
their products) best suits a rsi from the old country of 
Madhyadesa in ecstasies over his visit to the younger and 
developing Punjab settlements; and the simile drawn from 
Gandhara ewes betrays a non-Gandharan appreciation of their 
soft fleeces. 

Silk is more common in Vedic ritual use than woollens. 
Thus the ' vasas ' of ' tarpya,' 4 some sort of silk, 5 with which 
a dead body is clothed in order that the departed may go 
about properly dressed in Yama's realm, 6 was evidently an 
ancient traditional item of clothing; k$auma/ another 
variety of silk, is found early and often in ritual use 7 ; and 

1 Of. Pargiter: AIHT. chap. XXV. 

2 The Dharstas and Nariyaata Sakas were the first to settle in the 

Punjab j after_them came the Alias, in two nuin groups, 
Druhyus and Anavas, 

5 Just- as timber and bamboo styles of architecture were gradually 
transferred to stone. 

Av. XVIII, 4, SI; Taitt. Sam. II, 4, 11, 6; Mait. Sam- IV, 4, 3; 
Taitt. Bra. 13, 7 5 1; 7, 6, 4; Sat. Bra. V, 3, 5, 20 (worn by 
Jongs at sacrifices; the * rupani nisyutani ' on it show that it 
was something like mod. * kasida ' work on * tasar/ 

5 Most probably the sacred tasar,' a rough silk, the traditional 

product of E. Bihar. If the comm, has any basis for its 
explanation, * made from Troa or Triparna leaves.' these would 
refer to mulberry or other leaves suitable for silk-cocoons. (A 
variety of * tasar * (prob,=Chin. * tsau r and Burm. ' tsa ') 
produced in Bengal and Bihar is called * jarvo * or * jam * ; the 
habitat of ' tassr * is N.E. Deccan continued into Bengal aaid 
Bihar, and its hereditary growers are the Santals, with whom: 
it is a superstition and of religious and mysti significance : ci. 
Watt: CoMm, Prod, of Ind., V 1003 ff.) 

6 AT. XVin, 4, 31, * 

I Mait. Saip. Ill, 6, 7; Taitt, Sam. VI, 1, 1, 3; etc.; also in Sutras. 

( 61 ) 

even .saffron-coloured silken garments ( ' katisumbha- 
paridhana ') l were sacred. This comparative position of silk 

further explains and supports what has been said above. So 
also, garments made <of bark (so frequent in later literature) 
are very rarely mentioned in Vedic texts : and such and 
similar use of bark is more or less characteristic of the N. W. 
Himalayas; probably the ' barasi ' 2 of Kath. Sam. was a 
barken stuff. 3 

* References to weaving are very common from the Ev. 
onwards; ' vaya/ weaver, occurs often in B-v., as also various 
uses of the root * va/ 4 The special term * vaso-vaya * 5 shows 
that other * vayas ' had already arisen, who produced sundry 
piece-goods, other than the standard * vasas * or wearing 
cloth 6 ; besides, there were the female weavers, ' vayitri V and 
' siri ? s, 8 from very early times. 9 Technical terms connected 
with weaving, like ' otu * (woof, web), 10 * tantu ' (yarn, 
threads or other filaments) , n ' tantra * (warp, or loom) , 12 
' pracmatana r (forward-stretched web), 13 are already of 
frequent application in the Vedic texts; the ' veman ' (loom) 

1 Sankfa, Aran. XI, 4. 

2 Kith. Sam/ XV, 4; Pane. Bra. XVIII, 9, 6; XXI, 3, 4; the 

Kathakas were North-Western and sub- Himalayan; in these 
regions the Baras tree (a red-flowered rhododendron) is still 
fabled to yield cloths. 

s Or is * barasi J 'after all a variety of cotton 1 cf . 4 bairati * as such 
a variety known to Dacca weavers (vide Watt; Comm, Prod. 
of Ind., s.v. Cotton). 

4 Vide V.I., s.v. 'vaya ' and ' otn/ 

5 Vide ante. 

6 This distinction corresponds fairly with the later one bet. * taiiti ' 

and ' jola * in Bengal and Bihar : the former beine * vaso-vaya's 
only, the latter producing na'pkins, covers, upholstering stuff, 
etc. (Q. Is 4 jola * conn, with Tamil * jabaii '= cloth ? * j(jh) abli* 
in several vern. means * shabby clothes or rags.* Probably the 
Vedic names Jabala and Jabala mean "of a weaver ('jola') 
family '*, and perhaps place-names like Jabalpur or Jabli 
originallv signifi-ed "weaver settlement"), 

7 Pane, Bra. I, 8, 9; cf. Sat, Bra.j III, 1, 2, 13fL 

8 Bv. X, 71, 9, 

* sarigai/ embroidered fringe, which is probably connected with 
vern. * sari.' Probably the ' siri- amma* of Buddhist sculpture 
is the presiding genius of household weaving and handiworks 
( ' patnis ' wove or embroidered cloths for their husbands : cf . 
Av, XIV, 21, 51), hence of domestic prosperity, the original of 
the classical ( Sn ' (known from Sat. Bra. onwards.) 

10 Bv. VI, 9, 2.3; Av. XIV, 2, 51; Taitt. -Sam. VI, 1, 1, 4; etc. 

11 Av. XIV, 2, 51; cf. XV ; 3, 6 { -arob. = f gut ' ) ; Sat. Bra. HI, 1, 

2 f 18; Bv. X, 134, 5 (plant filaments). 

12 Bv, X, 71, ^ ; etc. 

13 Taitt Sam. VI, 1, 1, 4; 

{ 62 ) 

and fi mayukha' (peg, lead-weight, or shuttle) 1 are mentioned 
early In simile ; and the different parts of the ' vasas ' are des- 
cribed 2 in a manner that shows that it is the well-known cotton 
4 dhuti,' and presupposes a fully developed and long established 
indigenous cotton-industry, with which the \edic priesthood 
was quite familiar. It is to be noted that none of these and 
other terms, connected with * vasas J and weaving^ refer to 
woollen, or other manufactures ; where silks are intended, 
their specific names are given, like ' tarpya ' or ' ksanma ' ; 
and similarly woollens are distinguished as ' vasas/ derived 
from ' avi "s^or ' urna. 1 Thus the frequently used 4 general 
terms, * visas/ * vasana/ s vastra/ etc., with all their mani- 
fold parts and appliances for production so often detailed, 
can only refer to the Gangetie cotton manufactures, probably 
a prehistoric craft, with which the Vedic or Brahmapic 
civilization began. Accordingly we find the * vasas J being 
called sacred and divine in every part of it, in the ritualistic 
tests. 5 

The ' vasas * known to the average priest is practically 
of the same type in the several Samhitas and Brahmanas ; 
and its descriptions would apply equally to the modern hand- 
loom products of Bengal. Apart from its obvious analysis 
into threads constituting warp and woof (' otavah ' and 
* tantavab/ 6 or * otavah J and * pra-clnatana/ 7 or ' paryasa * 
and * anuchada ss ) , it had borders and fringes and ornamental 
embroideries, for which a number of technical terms are given, 
showing the same variety and importance of these in Vedic as 
In later times. Thus * sic ' is a general term 9 for the sewn on 


1 Vaja. *?am. XIX, 80; 83 (intermingling of liquors like shuttle 
through the loom). Hayukha=Maku (shuttle) of the Bengal 

* E.g. BY. I, 95, 7; AY. XIV, 2, 51; Taitt. Sam. VI, 1, 1, 3ff; Ka^h. 

gam. XXIII;, 1 } gat. Bra. Ill, I, 2, 1348 ; etc. 

* Probably * nttuda,* in Ay. Ill, 25, 1, means "sprung from * tuda ' 

or mulberry," i.e. ' silken ' (coverlet). 

* 4 Vasas ' : Rr. "l, 34, 1 j 115. 4 ; VTCT, 3, 24 ; X, 102, 2 5 etc. ; Taitt. 

Sam. VI, 1, 9, 7; 11, 2; Vaja. Sam. II. 32; XI, 40 5 Ait. Bra. 

I. 3; pf c. ; s vasana s : Rv. I, 95, 7; Cliind. Upan. VEtt, 8, 5; 

Sans- Upan. II. 15; ( vastra' : 'Rv. T, 26, 1; 134, 4; III, 39, 

2? IV, 38, 5; V, 29, 15; etc.; Av. V, l t 3; IX, 5, 25; XH, 

3, 21 ; etc. 

5 E.g. Taitt. Sam. VI. 1, 1, 3ff. 

i Kg. In Av. XIV, 2. 51; cf. 1. 45; Kath. Sam. XXIH, 1. 
7 E.g. in Taitt. Sam. VT, 1, 1, Sff. 
a E.g. m Sat. Bra. in, 1, 2. 13ff. 

* Probably * himya 3 in Bv. I, 34, 1, is another such name for borders 

or fringes of a cloth, which are inseparable from it, or from one 
er (vasasak himyeva) ; if * himya * may be derived from 
' (cf. Sayana), it can be compared with 'praghata 1 ; also 
* a soroewnat later term for these: **.g. in Sat Bra. HI, 

3, 2, 9; strainers with *da4a*: Ait. Bra- VII, 32; Sat. Bra. 

IV, 2, 2, 11; 1, 1, 28; and In the Sutras, 

( 63 ) 

or embroidered border or fringe (corresponding to modern 

* par * and * ancla *) ; two such are sometimes specified, 1 
showing the same old style of having two lengthwise and two 
breadthwise borders (the latter being the & ancia *s) of the 
same design for each pair; where the child is covered by its 
mother* s ' sic,' 2 where a deer-horn is tied in the) sacrificer's 
4 sic/ 3 or where the horizons at sunrise and sunset are said 
to be the two ft sicau ' 4 of the sky-cloth, it is the breadthwise 
broader border: elsewhere it is the lengthwise narrower one, 
or all the borders. 5 This wider border (corresponding to the 

* ancla ') is specially designated the fnvi/ 6 the closely woven 
end of the cloth, from which depends the * praghata * (or 

* the strikers'), 6 the loose and long unwoven fringe with 
swaying tassels 7 ; the ' visas * had only one * nlvi 9 usually, 
as now, the other end of the cloth being much plainer 8 : to this 
plainer end would belong the * tusa * 9 (or ' the chaffs'), a 
shorter fringe (corresponding to modern * chili, s or ' chilka," 

* chaffs]). The * vatapana/ 10 mentioned in two passages 
descriptive of the s vasas 9 as part of it, obviously cannot mean 
4 a garment to protect against winds * : it is ratter that part of 
the cloth which protects it against winds, i.e., its lengthwise 
borders, 11 which keep the web together from becoming thread- 
bare by fluttering in the wind (specially during movements) . 
The ' Irokah >12 (or * the brilliants J ) seem to have been 
flowers , stars or other spotty patterns 13 embroidered all over 
the cloth (corresponding to modern ' phul,' ' buta/ etc.). 

1 E. g . BV. I, 95, 7. 

2 Bv. X, 18, ll=Av. XVIII, 3, 50, 

5 Sat. Bra. HI, 2 ? 1, 18. 

4 Bv. I, 95. 7. 

5 E.g. in Av. XIV, 2, 51; Bv. HI, 53, 2. 

6 Taitt. Sam. VI, 1, 1, 3f; Kath. Sam. XXOT, 1; Sat Bra. Ill, 1, 

2 a ioff . : * nivi * is probably from Tamil '" siev/ to weave ; cf. 
vern. " newar *= woven straps. 

? So also, tb^ * praghata * is dedicated to plants or serpejats: the 
* antah J of Av. XTV, 2. 51, Is clearly = c praghata.' 

8 Specially in the case of men's cloths, this end being tucked BB 

i i_ i 


9 Taitt, Sam. I 3 8, 1, 1; II, 4, 9, 1; VI, 1, 1, 3; Kath. Sam. XXHE, 

1; Taitt. Bra. I, 6, I, 8; Pane, Bra. XVH. 1; etc. The 
Vratyaa favoured braided * tusa * fringes (dama-tusa^i). Appa- 
rently the 'tantavah' of Sat. "Bra. EH, 1, 2, 13ff*=*tusa,' for 
there * otn * and ' tantu * are already represented by * paryasa * 
and * anuchada.' That c tusa '= chaff, like lashes, is sho-wn by 
its dedication to Agni. 

10 Taitt. Sam. VI. 1, 1, 3ff. ; * vatapa ' : Kath, Sam. XXIH, 1 not in 

Sat. Bra. HE, 1 3 2, 13ff.. where however * sic * occurs. 
^ Probably preserved in the * bat an * { = border) o! the Benga] 
weavers: e.g. in ' gola-batan * cloths (cf. (?) * battnaul-kai ' : a 
caste of Madura weavers): also in veru. * bata '= split -bamboo. 
used in ptrengthening borders of thatches, etc. 

12 Sat. Bra. 1H ? 1. 2, 13ff. ; * atirokah * : Kath. Sam. XXIII, 1 j 

6 atika^ah ' : ^Faitt-. Sam. VI, 1. 1, 3ff. (probably wrong reading). 
* Arokafe ' may be an adapted form of the Tamil * arnkani ' 
ornamental border of cloths, Cf. classification of shawls, 

as * ek-rokha * and * du-rokha * ace. to the nature of their 

broldered patterns- 

13 So ialso ; they are dedicated to the * naksatras, 3 

( 64 ) 

For ritual purposes the cloth had to be unbleached and 
unwashed, 1 but ordinarily it was worn white: ,as by the 
Vasisthas. 2 Dyed 3 cloths with rich gold-thread brocades were 
affected by gay young women (typified by the attire of "Osas) 4 ; 
and red and gold borders are indicated by their comparison with 
the horizons at sunrise and sunset. 5 But the Vratya 
4 gjhapati's favoured dark-blue (' krsnasa ' ; antelope-hued) 
cloths and borders. 6 

The manner of wearing the cloth is not directly indicated 
by any reference. The ' visas ' however is always * tied, 8 
' girt/ etc, ( A nah 5 ), 7 which implies tucks and knots. The 
idiom * nivim ky ' 8 shows that each individual wore the * nivi ' 

1 E.g. in gat. Bra. Ill, 1, 2, 13ff. ('ahata'). Cf. the traditional 

distinction in the uses of * kora ' and f dholai ' cloths. 

2 Rv. VII, 53, I (Svityaneah) j cf . 83, 8 ; cf. 4 sukram atkam ' ; 

Rv. 1, 95, 7; * niktara. atkam': IX, 69, 4. 

3 The female cloth-dyer ( ' raiayitri ' ) is known early: e.g., Vaja. 
Sam. XXX 3 12; Taitt. Bra. Ill, 4, 7, 1 

4 E.g. Rv. I, 92, 4; X, I 6. 

5 fcv. I, 95, 7. 

6 Pane. Bra. XVH, 14-16; cf. Kat, Sr. Sut. XXII, 4, etc.; a akrnam 

krsnadasam va ' is added in the Sutras, 'and the name * kadgu * 
(preserved in vern. ' khadi ' and 'khaddar? ') is given to these 
varieties of cloths ; ' valukantani J of the older passage is explained 
in the Sutras and comm. as red or blue-black borders, which is no 
explanation; as * damatiisani ' m the same phrasa denotes a 
* style ' of * tiisa \ l valiikantani ' must iceaii a * style ' cf the 
broader border, i.e., * falling in fofda', or ' pkated * (val-uka) ; 
(probably done tip with the help of 'gila 1 and 'sankh* in the same 
manner as * desf s dhutis In the fashionable Bengal zemindar *s 
wardrobe) , The * krsnasa, * vasas would correspond to the 
modern * nilambari ', which 'as well as bine-black borders 
(krsna-dasam) on an indigo-dyed web, are Bengal specialities 
and favourites. (For the descr. * antelope-hued ', cf. mod, 
1 peacock-throated ' variety) . (It is curious that this peculiarity 
of the Vratyas should in later days belong to the Mahomedan 
population of Bengal (also in some other provinces, e.g., Punjab), 
who affect- the blues as opposed to the reds, which 'are the 
4 Hindu s shades in weavers' tradition (cf. Watt: Comm. Prod. 
of Ind., s.v. Cotton) j in the same connection Vratyas are said 
to use * silver * ornaments instead of the usual gold ( naturally, 
as silver occurs chiefly with iron ores in which the ** Vratya, 
country " is rich) ; this, again, is a Mahomedan trait in Bengal. 
It may be noted that Bengal Mahomedans represent a large 
section of the indigenous basic population of the Province.) 

* B.R. Av. XIV, 2, 70. 

* AT. VHL 2 S 16 (what * nivi ' thou makest for thyself) ; 6, 20 

{2 herbs to Ibe borne in the woman's * nivi *, evidently afc th 

navel) ; XIV, 2, 49-50 (make thyself a * nivi ' of this ' vasas * 

where the context shows that the * mvi * is hanging folds of the 

* vasas 1 ). It is difficult to see how a separate inner garment 

ean be meant "by c nivi * in these passages (so V.I.) ; * nivi * 

is distinctly stated to be a *part of tne cloth, like 'praghata*, 

etc. In later use also * nivi * is a knot, gatlfer or tuck, at the 

navel, of the fringed border which is primarily the * nivi/ 

Thai, in those passages * nivi * ta best taken as the * styto ' of 

"( 65 ) 

(or ancla) in his or her own way : evidently this refers to the 
same styles as the elaborate pleats and artistic waist-knots 
(mvi-bandha) of men and women in the early sculptures and 
classical paintings and poetry. The ' nivi s thus represents 
the modern ' kofica ' (pleats) and ' gant ' (knot) 1 : there is 
however no trace of the ' kacha ' (tuck of the plainer end of a 
cloth at the back) . Probably the Dravidian style of wearing 
the cloth without such posterior tuck 2 was at one time the 
fashion in 1ST. India also. The Vratya preference of braided 
or tasselled ' tusa * fringes 3 may, however, indicate that, 
while others tucked up the * tusa * in a full gather, the 
Vratyas displayed the hanging ornamental fringe, by tucking 
only "one corner of it. 4 The * nivi '-knot was sometimes so 
fashioned as to form a pouch, wherein magic herbs could be 
borne. 5 Sometimes, also, the ' nivi ' consisted of simply two 
1 tuckings up J (' udgiihana 5 ) 5 at the sides (as now, specially- 
with men) . Elsewhere women are said to tie their ' nivi * on 
the right side of the hip, the * nivi * being then covered by the 
upper garment ; such * nivi * must have been an ample gather 
of folds and fringe-tassels, for there a bundle of * barhis * 
represents the * nivi.' 7 It seems probable that women did not 
wind a part of the * vasas * over the bosom and shoulders (as 
now generally done in N. India) , which covered only the lower 
half of the body (as in Malabar, etc.). The description of 
Usas wearing rich brocaded cloth, and yet displaying her 
bosom, 8 would suggest this latter .style ; the ' nivi * style itself 
implies that no part of the broad border was left for such 
covering, and the early sculptures, etc., do not show it. 9 
Apparently the upper part of the body of men and women was 

wearing the * nivi ' or border. It Is possible however to see In 
' yat te vasa^t paridhanam, yam nivim krnuse tvam % a 
reference to the ordinary * wearing cloth r and a separate 
specially woven strip to serve as an artistically tied waist-band, 
something like the * commerbund ' of medieval Dacca manu- 
facture; this specialization and separation of the * nivi ' is also 
shown in quite early sculptures, etc. But even in that case the 
* nivi ' would be an outer adjunct and not an * Inner garment, 
forming- one of three. * 

1 The former is more in evidence in Bengal, the latter in Bihar and 


2 Curiously, again, the Bengal Mahomedans affect this tuckless style; 

cf. n. 7, p. 64, (re Vratyas). 

3 Vide ante. 

4 Also a Dravidian peculiarity. 

5 As in the Bihari knot. Av. VHI, 6, 20. This style cannot have 

been confined to women, as not much later on, * nivi * came -to 
mean deposit money or capital, 

gat. Bra. HI, 2, 1, 15. 

1 Sat. Bra. I, 3, 3, 6. 
8 Bv. I, 92, 4. 

S In these the ftpper part of the body is often bare, covered only by 
various elaborate ornaments: sometimes 9, few lines are indicate^ 
to show a filmy wrap. 

( 66 ) 

covered when necessary, by another separate garment either 

- ISS,iike ' upavasana,' ' paryanaham/ or adhivasa, 
iSTilSinaaa clle-fittmg jacket, bodice or cloak like the 
pratidhi/ 'drapi,' or ' atka. 1 Thus the bride had her 

* upaTaaam, 9 apparently a scarf or veil 1 (corresponding tp the 
modern ' orna ' used by women), an,d the vasas ot 
MudaaSni that fluttered high up in the air, 2 was evidently eueh 
an 'uttariya' scarf. ' Soma, 1 in the ritual has _ his 
4 paryanahana; in addition to his 'upanahana and tisniip 
from" which a strip two or three inches wide might 
be torn to form an ' usnisa,' if necessary ; so that 
the ' naryanahana ' (lit. wrapped round about) was a 
pretty long and ample scarf of ligiht texture.* The 
* adhlvasa ' does not seem to have been close fitting like 
the ' <atka ' or * drapi,' 6 as it is an ' over-garment/ worn by 
princes over their inner and outer garments 6 ; again the 
forests are the ' adhivasa ' of mother earth licked by the fire- 
child 7 ; it was thus more like a long loose-flowing dressing-gown, 
suiting both men and women 8 ; it may not,, however, have been 
a tailor-made garment at all, being called a * vasas ' 9 ; probably 
it was of the same sort as the ' upavsLsana.* The * pratidhi ' 
must, from the contest, 10 refer to a part of the bride's attire, 
apart from the newly woven, excellent garment 11 ; apparently it 
consisted of one or two strips of specially made cloth drawn 
across or crosswise over the bust and tied at the back, to serve 
as a bodice, 12 or was a short and tight bust-bodice like the later 
* kanculika * (mod. * kaftculi ' ) . The ' drapi * seems to Shave 
been a close-fitting 13 and gold-embroidered 14 vest, 15 used equally 

1 Av. XIV, 2, 49 and 65. (In the latter passage it may mean coverlet 

of a couch, being mentioned along with furniture). 

2 Rv. X, 102, 2. 

3 These three may well be rendered by the mod. terms, * cadar ' (or 

uniiii), * dhuti ' and " pagri/ respectively. 

4 gat. Bri III. 3, 2, 3. 

5 So V.I. 

B Sat. Bra. V, 4, 4, 3. 

7 By. I 3 140, 9. 

8 "Vide it 7 above (matuh) ; cf. Kv. X, 5, 4. 
S E.g. Rv. I. 162, 16. 

1Q Av. XIV, 1 8. 

11 Av. XTV, i f 7,45 ; the usual refer enca to a part of the chariot ia 

hardly 'appropriate. 

12 This style is now found amongst Kolarian races, and is a specially 

festive one. (Of. the cross cords in Hellenic drapery). 
B Rv. I, 166, 10 (Cyavana's old age like a * drapi ') j probably 
* drapi *=a tight vest suitable for running about (dra). 

14 RY. I, 25, 13 (Mranyayam) j IV, 53, 2 (pisangain) ; Av. V, 7, 10 

(* hiraaya-drapV adj. of a woman). 

15 ?Av. XHI, 3 t 1 (the sun wearing the 3 worlds, making a * drapi * 

of them : hence the * drapi T had three pieces, two side ones 
and one back, like a waistcoat ; it was not a * coat of mail * (so 
V.I.) being worn by women, and the se of r vasanah," etc. 
(cf . * drapiio, vasanah/ Rv. IX, 86, 14) would rather show 
it was made of * vasas * 

( .67 ) 

t>y men and women, 1 specially tsy prominent men 2 and gay 
women. 3 The ' atka ' was confined to men ; and was a long 4 
and fully covering, 5 close-fitting 6 cloak, bright 7 and beautiful/ 
the .stuff being bleached 7 cotton, 8 interwoven 9 or embroidered 10 
with gold threads. s Pesas J is gold-embroidered cloth gener- 
ally 11 ; the designs were apparently artistic and inttieafe, 12 and 
the inlay of gold heavy and brilliant 13 ; where, however, the 

* nrtu * appears with * pesamsi s on, M it might refer to a pleated 
skirt made of such brocaded cloth, like the medieval and 
modern * ghaghra J or * peswaz.' 15 . It is noteworthy that the 
early Vedic references to ' atka/ pesas/ e samnlya, * and 

* drapi J come mostly from Angirasa poets 16 ; these were 
therefore primarily East Indian styles. Curiously enough, 

1 Of. 'the same style in N. W. India, where both men and women 

show of their richly embroidered -waistcoats. 

2 Bv. IX,, 100, 9 (wearing * drapi ' on becoming great). 

3 Av. V, 1, 10 ('hiranya-drapi' worn by Arati* likened to a 


4 Bv. n, 35, 14 (food carried in one's own c atka ' : Le., in the long 

skirt made into an apron). 

Bv. V, 74, 5 (' vavrim atkam, likened to Cyavana's old age : prob- 
ably being a tight fitting garment it showed many creases 
resembling wrinkled skin); c. IV, 18, 5 (Indra Bora with 
* atka '= his own covering glory). 

6 'Surabhim atkam': Bv. VI, 29, 3; X, 123, 7. 

7 'Like snn' : Bv.* VI, 29, 3; X, 123, 7; ' aTjuna ': fj- I*, 10 J'Jf ; 

' snkram ? : Rv. I, 95, 7 j ( niktam ' : Bv IX, 69, 4 ; sudrsi : 

T> T 1[02 9 

8 As ' vyiitam ' 'and frequent nse of ' vasanah ' shows ; it cannot very 

well have been an annonr (as sometimes translated and 

explained) . 
8 * Hiranyair vyiitani ' : Bv. I, 122, 2. 

10 ' Hiranyayan ' : Bv. V, 55, 6. , 

11 Bv IV, 36. 7 (the best and attractive * pesas ' spread for the gods) ; 

"of. < hiranya-pes'as ' worn by a hoiise-holder and his ; wife: 
Bv Vni, 31, 8; VII, 42, 1; Vaja. Sam. XIX, 82; 83; 89; eta 



12 Bv f S 3 eTcrVajaSan;. XX, 41 (design compared with the 

*Doets''son0. The manner of ' pesas ' work described here is 

the same L the ' jari ' and < salma-cumki ' work in the present 

^11 34 11 (the glittering surf a- of rivers = ' pesas '-Varnna : 
the * iarr work if most faithfully described in this pas sage; 
one who has seen the J** 


Ganges will appreciate the similarn. . . 

pesos' is 'apparently called bright as ( ghee Mi.e golden). 
- T 92 4-5- cf. also * yuvatih siipesah : Kv. A, IIH, o. 

freqnently in Upper India, but are specially associated 

* Thuf -W^? SfaB- 5; ^T. 29, 3 5 VIII 
ThU ai 7 are the , Tr^r'asa rrfs. : 3. other rtfc. are Xtteya d 2 
Bharrav re ' pesas.' the Angirasa ref s. are : Ev I. 92, *, 
TV K .7'; 2 others being VBiirtha and Bbarpva : re satnya , 
ihe only early Vedi, reference is in Av. (ocenrnng 

another Bliargava. 

( 88 ) 

these are preserved in the later ' ackan/ 1 * peswa (z),' 2 and 
* samla/ 3 winch agree fully with the Vedic items of dress ; they 
are usually supposed to be derived from the Persian ; but nior 
probably it is a case of re-imposition of Persian stamp upon 
common Indo-Iranian items of material civilization; in fact, 
the Persians must have ultimately derived these styles from 
their Western- Aila ancestors, from the Puranic point of 
view. 4 

It is remarkable that the * usnisa ' is not mentioned in 

early Vedic literature, except in connection with the Vratyas 
in the Av. 5 ; it appears, however, oftener in the Yv. Samhitas 
and Brahmanas, but again chiefly in connection with the 
Vratyas 5 and Kings/ It seems likely therefore that turbans 
were" not originally in use, 8 and were introduced as a style 
through the Vratyas of the Praci, amongst whom kingship is 
said to have arisen. 9 The Vratya's * usnisa ' was bright and 
white as day (while his hair was dark as night) 10 : it was 
evidently of some fine cotton stuff 11 ; this was (according to the 
Sutras) tied with a tilt and cross-windings. 12 The King's 
'usnisa* was tied in a special manner at ceremonial sacrifices 13 : 
the ends were gathered together and tucked away in front, so as 
to cover them up, 14 this tuck at front being preferred by the 
Sat. Bra. to the other ritual style of winding the turban quite 

1 c Ackan* nsed to be an item of respectable Hindu dress (as opposed 

to Mahomedaa), but is now used chiefly by waiters or menials 
in Anglo-Indian establishments. 

2 *Peswa '=women j s garment; * peswaz ' =full-dress gown, sp. of 

dancers. Such special dancers' dress was noted by Greek writers 
of the 4th cent. B.C. 

3 Tide p. 59 ; JL 6. 

4 Vide infra, sec. re Persian influence in early social customs. 

5 Av. XV, 2, Iff. (where it is one of the ' characteristics * of the 

Vratya chieftain). 

6 Pane, Bra. XVI. 6, 13; XVII, 1, 14 (amongst ' vratyadhanani ' ; 

read * Mhanani *?) 

7 Mait. Sam. IV, 4, 3, etc. (ksatra at sacrifices); gat. Bra. HI, 3, 2, 3 

(King ' Soma ') ; V, 3, 5, 23 (King at sacrifices) j XIV, 2, 1, 18 
(Indrani); etc.; (fee of gold presented in an 'usnisa', in 
3 fcosas: Kath. Sam. SHI, 10; Taitt. Sam. 133, 4, I, 4*)." 

8 The only head-dress known to Ev. being the c sipra/ a sort of helmet, 

evidently used only in battle: e.g. .Rv. V, 54, 11; VIU, 
7, 25 ; etc. Probably the Aryan incomers wore felt caps and 
hats (like various Scythic or Iranic tribes). ' Stupa,' in the 
loosened * stiipa * of Aryaman, or in the proper name * Hiranya- 
stupa/ may mean the Vedic ' topi ' (' tupi ') or conical cap; "for 
the shape, cf. the tiaditional ceremonial < cap % * topara ', 
resembling a ' stiipa ' structure. 
8 0$. AV. XV, 2-10, which agrees fully with the unanimous Puranic 

s * ^S^^ 011 7e ^ rst kmgs i n the Siita-Magadha country. 

30 Av. XV,, 2. 

11 Like the muslin * pagri-doth * traditionally used. 

12 ' Tiryan-naddham ' ; Kat. Sr. Sat. XXI, 4. This is the traditional 

style again. 

fiat. Bra. V, 3, 5, 20fl 
14 * Samhrtya purastad avaguhyati \ 

( 69 ). 

round about. 1 These special styles show that ordinarily the 
princes wore turbans with loose hanging ends, 2 which were 
inconvenient and dangerous in ritual ; accordingly, elsewhere 
in ritual, the ' usnisa ' is only a kerchief 3 ,: probably this 
kerchief tied round the head (in TibetoJBurman or Kolarian 
fashion) was the original brahinanical ' usnisa/ so that 
when ruling princes joined in their rituals, they had to adopt 
a trimmer form of their unwieldy turbans 4 ; so also Indian! f 
wears an * usnisa ' like a zone, of variegated hue/ clearly a 
head-band of a many-coloured silken kerchief. 6 

No general footwear, again, is mentioned in the earlier 
Samhitas. 7 ' Padvisa ' 8 in the Kv. is applied to the leggings 
of ahorse 9 ; Vaturina pada/ 10 probably refers to heavy frnaha') 
covering footguards, used by chiefs 11 in battle ; ' pat-sangini ' 12 
in the Av. also refers to somewhat clumsy hampering foot- 
fasteners used by soldiers. 13 The k upanah ' first occurs in the 
Yv. Samhitas 14 and the Brahmanas, as used in ritual 15 and by 

1 There is no mention in the Bra. text of the ends of the turban 

being tied behind, drawn over one shoulder like an * upavita ', 
and tucking in the waist-cloth. (Of. Eggeling's note in S.B.E. 
and comm. on the passage). 

2 In traditional style: e.g. in Upper India generally 9 specially amongst 

military castes. 

3 gat. Bra. IV, 5, 2, 2. 7j the * usnisa ' that is tied round the eyes 

of the ' naga seer ' Arbuda, seems also to be a kerchief only : 
Ait. Bra. VI, 1. In Sat. Bra. HI, 3, 2, 3, the ' emergency ' 
turban bonnd with a strip of cloth 2/3 Inches only in \vid*-h, 
shows that the turban was often a mere band, or a ' ropy * head- 
gear with many twists, like that affected by Deccanis. 
3 Cf . the modifications of the. royal ' asandl * in ritual. 

5 Sat. Bra. XIV, 2, 1, 8. Of. Ham. VI. 80 (" Kak." women wearing 

red * usnias * whild assisting at Indraji&'s sacrifice). 

6 A Kolarian and Burmese feminine style, again. 

7 Tradition however ascribes a high antiquity to tie * upanah * (and 

the ' chatra J ) : it is said that Jamadagni-Bliargava (contemporary 
of the famous < Rgvedic Visvamitra) Introduced their use for the 
comfort of his delicate wife^ the Aiksvaka princess Benuka 
(cf . Mbh. XIII. 95) ; hence either the Iksvakus took to shoes 
and sunshades after the Bhrgus^ or, more probably, the Bhrgu 
brahmans learnt their use from the * Solar * court* after Jama- 
dagnfs politic marriage. Ifc is curious that the Av. (VI, 156-7) 
ascribes the first preparation of a potent hair-tonic to this 
Jamadagni-Bhargava, who prescribed it for his daughter (ap- 
parently as stylish a lady as her royal mother !). All this agrees 
with the well known fact that the Bhrgus were the most 
" Ksatrlyanised " of the brahmans. Probably other brahmamc 
groups were not accustomed to shoes, etc., till later on: hence 
there is no very early mention of these in priestly literature. 

8 Bv. I, 166, 16. 

9 In Av., to foot-fetters or shacklea (VIII, 1, 4; XII, 5, 15, etc.). 

10 Ev. I, 133, 2. 

11 Indra crushes enenay heads with them. It Is Interesting to compare 

the Indo-Scythic boots, as in Kanlska's statue- land Kushan coins. 
!2 Av. V, 21, 10. 

13 Apparently worn only during long marches or rapid flights. 

14 A "so In Av. SX, 133, 4, a late passage. 

15 Taitfc. Sam. V, 4, 4, 4; 6, 6, 1, etc., Sat, Bra. V, 4, 3, W $ Kati$. 

Bra. Ill, 3 (staff and sandals). 


( 70 X 

the Vratyas. 1 The ritual sandals ctf shoes were made of black- 
antelope- or boar-skins 2 ; those of the Vratyas are described 
in the Sutras 3 as black and pointed C'), etc. 4 ; these 
details indicate that the most stylish shoe-wearers of those days 
were the Vratyas, just as they were the chief wearers of the 
4 usnisa.' It is probable that the use of footwear in early 
times 'was to some extent limited by the common fashion 
(with both men and women) of wearing ' khadis ' or anklets. 5 
Similarly the use of the ' usnisa/ also, must have been 
restricted by the preTalent fashions of hair-dressing. Whole 
clans had distinctive styles of wearing the hair : thus the 
Vasisthas could be recognized by their white clothes -and 
* kaparda ' worn on the right side of the head 6 ; so that they 
could never have used turbans; and (as already noticed) 
apparently no brahmans originally used them. Another style 
of hair-dressing was wearing the ' kaparda * in front 
(* pulasti y ) 7 ; it seems probable that the Pulastyas (an early 
brahman group cognate to the Agastyas, and like them asso- 
ciated with Deccan non- Aryans) 8 were so-called from this 
distinctive style, * Kesara-prabandhayah ' in the corrupt 
Atharvavedic passage 9 yields much better sense if read 
iu -prabandhanam' (specially in view of the fact that the whole 
context 10 refers to tihe results of the famous Haihaya-Bhargava 

1 Pane. Bra. XVII, 14-16. 

2 Vide note 15, page 69. 

3 E.g. Kat. fir. Sat. XXII, 4. 

4 According to details ia other Sutras (.and comm.), these were also 

variegated, or like ' varma,' i.e.. with metal knobs; etc, 

5 Gf. Ev. V, 54, 11; etc. 

6 Bv. VII 3 33, 1; 85, 8. 

7 Vaja. Sam. XVI, 43 ; it is usually taken as meaning * wearing the 

hair plain ' ; but * placed in front ' suits the context much better, 
for Eudra's 'kaparda ' is traditionally inclined in front (' pulasti- 
kapardia *). It seems however equally probable that this 
* pulasti ' style was so called being that affected by the Pulaa- 
tyas, and not for being a frontal mode of ' Kaparda '-dressing. 
The clan-names of the Pulastyas, Pulahas and Agastyas mean 
the same thing (previous inhabitants), and correspond to the 
Pelasgu of Hellenic history; together with the Kratus they seem 
(from Puranic evidence) to have formed an earlier (pre-Aryan) 
stratum of civilization in India ; they were finally either absorbed 
after struggles with Manvas and Ailas, or were expelled sea- 
wards and westwards. 

8 In all Puranas; the Pulastyas would thus seem to have been 

^Saivites." (Q. Has 4 kapardin/ etc., a phallic symbolic 
s ; giiincance? i.e., from * kaprth ' ; in that case the Yas'isthaa 

and Pulastyas may have b&en ' phallic * priests originally). " 
s Av. V, 18, 11; the various interpretations of commentators and 
translators making kesara-prabandha ' a cow or a woman with 
a she-goat, etc.,, are absurd; if the passage has to be emended, 
the above emendation (with e caramajan * for * caramajawn * etc.) 
would be best: the sense would then be, that the Vaitahavyas 
who destroyed even the new-born babes of the ' kesara-prabandhafr ' 

Bhrgus, perished with their whole kin, etc-, which in fact is 

the unanimous tradition. <r 

10 Av. V. 18-19. This is a very remarkable early brahmanical version 
of the famous Haihaya raids of Puranic tradition. 

X 71 ) 

conflict) ; this reading would show the Bhargavas to have been 
' kesara-prabandhah/ or ' wearing braided hair like manes/ 
quite in agreement with similar Yedic references to brahmani- 
cal hair-dressings. Some of the Vedic gods wear * kaparda' s 
and l opasa J s ? apart from goddesses like Sinivali 1 .: thus Eudra 
has his hair in the * kaparda ' style 2 ; so also Puan?; and 
India's * opasa * is likened to the vault of heaven. 4 These 
divinities have marked indigenous and extra-" aryan ' features ; 
and it is .significant that peculiar styles of hair-dressing to the 
exclusion pi" those of head-dresses should be characteristic of 
brahma^s and such gods. 

Women of course wore their hair in a number of different 
styles, which are, however, rather vaguely indicated by the 
special terms, 5 ' stuka/ 5 ' kurlra/ or k kumba,' besides 
the ' opasa ' and * kaparda ' mentioned above. In the 
first place, it seems clear enough that b opasa ' and 

* kaparda/ being ascribed to men as well, were not dis- 
tinctively feminine styles, and could be managed' by the average 
long-haired man. Accordingly young maidens are said to 
wear their hair in four 4 kaparda's. 7 What the * kaparda ' 
of men was like, can be very well made out from the traditional 
representations of the * kapardin ' god and the hah' -dressing 
of his followers 8 ; it was a spiral coil of the braided, plaited or 
matted hair, piled on the top of the head at different angles. 
It was apparently the same in the case of women, for the 
maidens 1 four ' kaparda's are compared to the four corners of 
the altar, 9 and so cannot mean * braids 7 or ' plaits/ 10 while 
Sinivalf s * kaparda ' is an alternative style classed with 

* kurira * and * opasa/ 11 The four * kaparda's of maidens 

t It Is to be rioted that in Epic-Puranic mythology, Sinivali and 

cognate goddesses are specially Aiigirasa and domestic ones. 
(So also a chief feature of Indo-Aryan mythology is aBsence or 
unimportance of goddesses). Mudgala of Paiicaia, who became 
an Angirasa, wore a ' kaparda J (carrying afc the same time an 
* atra ' like Vratya chiefs) : Ev. X, 103, 8. 

% Ev. I, 114, 1, 5; Vaja. Sain. XVI, 10; 29 j 43; 48 j 59, (Occasionally 
a Budra, wears scattered tufts or has a shaven head : Vaja. 
Sam. XVI, 59 and 29 respectively). 

3 Ev. VI, 55, 2; IX, 67, 11. 

4 Bv. I, 173, 6; VIII, 14, 5; the sense of 'diadem* is not at all 


5 The commentators are hopelessly contradictory and evasive with 

regard to these terms. 

6 Gf , E. vern. * thoka ' =lump. 

7 Rv. X, 114, 3. 

8 The Saiva devotees; this style is also afiected by men ia Orissa and 

the S.E.j even now. 

9 Vide ante. * 
10 As taken in VI. 

TT- * ci -v j *"*- 

72 ) 

must have together f orated .a crown-shaped 'ooiffuro. 
The ' opasa ' as worn by men probably consisted in 
gathering up all the. hiair with a small top-knot, 
leaving it loose enough to fonn a dome -like cover or flounced 
cap ; this would .explain most of the figures in the texts con- 
nected with ' opasa ' : thus the ' opasa J s of Indra and Soma 1 
are like the clouded or vaulted sky ; the thatched net-covered? 
roofing of a house (compared to a woman) is like ' opasa ' 
spread over the * visuvant ' 3 ; and the knob -like horns of the 
year-old cow are ' opasas.' 4 These last similes show that the 
' opasa ' was of the same style in the case of women also, 
unless the qualification ' su * in Sinivali's description 5 is taken 
to mean a heavier * kaparda,' and an ampler * opasa,' and 
with the probable exception of the covering and withholding 
net 6 ; but * harih * * opasa- ' of Soma 7 might refer to coloured 
covering-nets used by men as well. It seems that sometimes 

* opasa (by a common figure) meant this covering-net only, 
as in the case of the bride's hair being dressed into a 
' kurira * and * opasa/ where the two apparently form parts of 
one composite coiffure. 8 The practical identity of the 
masculine sand feminine ' opasa 's is also sh^wn by the 
Av. charm, which regards the imsexing of >a rival as complete 
only when, after the ' opasa/ the * kurira ' and then the 

* kumba * are, in addition placed on his head. 9 These two 
therefore were the idistinctively womanly styles 10 : and they 
are, accordingly npt ascribed to men in the texts. As 

* kurlrin ' is used secondarily of a horned animal, 11 the 
' kurira ' must have been a horn-shaped coiffure, possible 
only with the long braids of women; a net' or veil 
( opasa/ 2 ,) may have been hung from this * horn. 12 The 

73, 6; VIII, 14, 5, and IX, 71, 1, respectively. 
(-eyed aku. 

3 Av IX, 3, 8,- the parting of the hair would naturally be covered 

by such cap-like ' opasa.' (It will be noted that such ' opasa ' 
would have a frontal aspect exactly like a curved thatch with 
hanging eaves). 

4 ^nc Bra. jy i 1; cf . < dvy-opas'aV in XIII, 4, 3, (It is not 

L iFi, 7 t. gr0 ^ n J 10 3 of kine that are referred to; the 
s TV*? T T?? n ? er f J ! vl(rasl y metaphorical and secondary). 

5 Twt J- S* I*, 1, 5, 3; Mait. Sam. II, 7? 5 5 Vaja. Sam. XI 56. 

i.2 Ti ? -! ee ^Y .' ? u ' can refer to a Vedio "custom of 
wearing false plaits of hair). 

ft Indicated by the simile in. Av. IX, 3, 8. 

With this may be compared the zone-like head-band of variegated 
^ toe worn by Indrani. Vide ante. 

9 Av* VI^TO 5 ^%^" ^ J ^' ^" ^^ e i n fra, r 

10 uf. Apast. Sr. Sut,, e< Kumba and knrTra nn 

11 Air T7 T1 1 / i * " fja( *.u ivuxna uu uiA^? ^aiv*ii o ncci 

horn) ' alre& ay noted, c opasa * cannot mean such 

13 As ?*Myw the ^bridrt hair-dress; (vide not<f8 above). This style 
* I G m hil1 tracts between ^ " ^ ' 


( kumba n Is evidently the vern. * khompa ' 2 of later times, 
the specially feminine, hemispherical or pot-shaped coil at the 
back of the head. 

1 The form and sense of the word suggests a connection with 
' kttmbha/ ' kambu,* etc., all implying something _ rounded. 
(Probably ' kumbya ! as a form of measured speech like saman 
or gatha is taken from some process or feature in .the ' kumba-' 
dressing : cf. Sat. Bra. XI, 5, 7, 10). 

2 Ifc is to be noted that the ' kumba ' occurs only in Av, (and^mucfa 
later on in Sutras) j the presumption therefore is that it was 
primarily an Angirasa, style; it may be connected with Tamil 
' kudum ' = ' coil of hair' and ' pa/ to weave or 'braid.' Cf. 
vern. ' kadam (ba),' a flower, and * kadma/ a sort of toffee, 
bofch obviously deriving their names 'from the various elaborate 
modes of the * kumba. 1 





There is no explicit statement in the earlier Samhitas (as 
there is in the Epic-Paranic literature') 1 of any notion that 
at & remote period the regular and correct marriage was 
unknown, and that the institution was gradually developed or 
introduced by way of reform. But there are clear indications 
in them that an established standard of marriage was only 
evolved through various preceding stages of sexual relationships, 
more or less primitive in character. 

Such relationship was not uncommon in early times as 

between brothers and sisters. Though it seems frpm the 

Yama-Yami dialogue 2 (which is best regarded as an example 

of a very early form of ' social driarna ') that, at the time 

when it was composed (apparently the latter part of the 

5gyedic period) such connections were coming to be regarded 

as incestuous, yet the very fact that this could be made the 

subject of a serious piece of composition with a ' moral ' in 

it, shows that they were still not very rare; thus Yama (an 

early legendary hero selected for effective illustration) is made 

to say lt verily there will come other ages wherein brothers 

will unite with sisters," etc., obviously referring to the 

practices current in the poet's generation, or at least those 

within the memory of his tinies. This is confirmed by other 

references 3 of the same or earlier period, which can only mean 

that brother-sister connections and wooings were quite normal 

and recognized, in the Vedic priestly society at least; thus, a 

favourite god is appreciated for wooing his sister ; the brother 

is classed with the husband or the paramour as a person 

normally approach-ing a woman ; and for the .sake of a son 

md heir, men may unite with their sisters ; while in one of 

the \edie marriage mantras 4 union with an adorned ' iami * 

(sister) sitting among the fathers, is regarded as Visvavasu's 

birth-iight, so that the contest would suggest that the mar- 

riage being celebrated was also one between a * ami ' and 

,( 75 }. 

her brother. 1 It seems probable, from the selection of Yama 
and Yarn! as a type (in the above poem), and from Yami's 
arguments, that twins were regarded in a superstitions primi- 
tive age as specially destined* for such relationship, more 
than ether brothers and sisters. 

The case of fathers and daughters is not equally clear; 
it is very early recognized as incestuous, but seems to have 
been once frequent, almost a permitted practice. Most of 
the references in the early Samhitas 3 to this form of incestuous 
connexion are explained mythologically in the Brahmanas.* 1 
Still the fact remains that such a relationship serves as a 
simile or lallegory, and is described in a manner that shows 
approval; and even alleged mythological features very often 
have a basis in primitive conditions, which the believers in 
those legends may have outgrown, or grow out of actual and 
traditional early events, to justify which legends are inter- 
woven in course of time. 5 The ascription of such connexions 
to Prajapati and his daughter or Pusan .and his mother, 6 .shows 
that the Vedic priest could still conceive of such relationships 
as not at rail damaging to the prestige of his gods. But actual 
amours of this type were known : thus there is a* plain reference 
to father-daughter connexions in the Av., which would show 
that these were common enough to be alluded to, and even 
presumed, in a domestic rite concerning women 7 ; while in the 
Ait. Bra. a very old * gatha ' is cited (in connection with the 
royal consecration and the Sunahsepa story) / where for the 
sake of sons men are said to unite with their mother and sister 

1 Vide infra, for the composite character of the marriage hymns. 

2 In Puranic tradition also, the twins of Ut&ara-Kuru are devoted 

married pairs all through life. 

3 E.g. Rv. X, 61, 5-7. 

4 Ait Bra. Ill, 53 ? 5 j gat. Bra. I, 7, 4, 1 ; Pane, Bra. VIII, 2, 10. 

5 Cf. the legend of- Yrsanasva's daughter Mena (Rv. I, 51, 13 and in 

Bras); Indra's applauded part in it seems to have been intro- 
duced to gloss over or justify an ancient brother -sister connexion 
(Indra plays a similar part in other legends of questionable 
morality) ; the Puranic inclusion of Mena in the * pitr-kanya ' 
group (vide infra) apparently presupposes such a tradition of her 
incestuous connexion. 

6 Rv. VI, 55 3 5; (also Pusan and his sister; ibid. 4). 

7 Av. VIII, 6, 7. The Atharva-vedic charms are mostly the products 

of stages of civilization earlier than the $gvedic; but this 
particular one being included in part within the Rgveda^also. 
must have represented more or less contemporary conditions; 
such conditions are regarded as normal amongst townspeople 
in the Jatakas; vide infra. 

8 Ait. Bra. VII, 15 ; cf . gaakh. r. Sut. XV, 17-25. This gatha * 

belongs at least to the 10th century B.C., while it refers^to 
Hariscandra's time, about eight centuries before that according 
to Purar^c tradition. For Puranic notices of incestuous unions in 
Aiksvaka and other dynasties of that age, vide infra. The practice 
in this * gatha ' however is advocated by rsis, before a rajanyst. 

{ 76 ) 

as with a wife, 1 Such facts probably point to the ultimate 
origin of the practice 2 of ' appointing ' a daughter to bear a 
Bon for the father, while remaining with him, such a son being 
regarded as the father's own son. 3 The father-daughter con- 
nexion 3 as a more or less recognized practice or permissible 
license., could possibly have originated in a primitive, strongly 
patriarchal group , which, being still unsettled and raiding 
about, would .at the same time have a minimum .supply of 
women 4 ; in such a case the -essential sons, 5 not ordinarily 
obtained, would come through the daughter. It is notewortty 
that * duhitr ' primarily implies no connection with the father 
as such, but simply denotes woman as * nourishes of a child f 
or ' potential mother.' 6 With the passing of primitive condi- 
tions the daughter's position would change, and she would 
come to be " appointed " to bear a son for her father's family 
in <an indirect way. 

It does not clearly appear how far the practice of sister- 
marriage was the result of a similarly strong patriarchal and 
isolative tendency or that of an earlier matriarchal state of 
.society amongst some at least of the Vedic tribes 7 : thus Yaml's 
insistence on the point 8 that Yama'-s conduct is unbrotherly, 
and for the sake of & protection * and * offspring s ( ' a grand- 
son for their father ') he should be her husband, is a patriarchal 
trait ; on the other hand her marked initiative in the matter 
and bold wooing 9 is a matriarchal one, while the position, is 
reversed in other cases. 10 The probability of the former condi- 

I The practice prevailed in ancient, Lrarj (an important point, since 
Puranic traditioii regards ' Iranic Aryans as subsequent off- 

\r2 4-L ? Ildland J Ar ? ans of India, amongst whom the 
Madras, Vahtekas and other North Westerners had similar 
practices; cf. Mbh. YIII, 40; 4V5); also amongst the old 

"vide g (IVj 5> 4)< 

rnMtrrn -+- i J 1 - -, " t0 klllt at tllat -v, u 

t^-'iralto'^^^o^to ^^ etc -' ) '- The technical 
3 ^tiQns]" SSt ' XXVIIIj ^ ( a s ^ tra P^serving muchof "oldeToMidi- 


to be regarded 
certain circles 

_ ^FLift J.A *" |r . *' * "*"v v jJoiienAcio mnaj. i. 

I i B Desire lor many sons is a most prominent early Vedic feature. 


t B is to be remembered 4hat (according to tradition) some of the 
vetiic tribes were originally Dravi^ian (non-Aila), and they may 
erenih * B"h mr - feat ^res late into the B,gvedic or 

ft irfrV ^C. 1 10 "f 1 ^*C 
10 BV - ILit It'l^ti^tfar; 601 " 6 f -*" 6i8ker - 

TFlOTSff ^TUT r% ^TtaC" r- .ft..-r_in i _ t T T w "8 

polyandnc, or show 

( 77 ) 

tions is suggested by facts like these": The e bhratr ' is not 
characterized by blood-relationship, but is primarily the * sup- 
porter and master } of the * svasr s and others 1 ; an external 
woman could be taken into a familv as a * svasr/ a sort of 

\P * * 

* companion/ 2 to be thus supported; sister as a blood-relation 
(' jami-svasa,' and then only ' jami f ) is -a later development, 
as shown by the adjectival use of * jami ' 3 ; * jnati ' means 

* brother and sister ' primarily, as being best acquainted with 
one another 4 ; the disputed precedence (referred to in a Bra .) 
at ceremonial family meals, of the sister over the wife, 5 pro- 
bably points to a time when the sister had actually the 
place of the wife, in the family and its ritual; when 
the sister was no longer normally in that position, she 
was still supported and controlled (in her social and marriage 
relations) by the * bhratr/ 6 On the other hand, there is 
some trace of the greater importance of the sister and the 
mother in earlier times: The sister's claim to precedence 

over the wife in family ritual is one indication 7 ; the dread of 

the * sister's curse ' 8 shows her early influence, probably as 
the original mistress of the family ; she is the best ' jfiati ' of 
the brother 9 ; 'the * putra ' 10 belonged specially to the ' matr/ 11 
and the mother is sometimes the * bhartri * or supporter of the 
family 12 ; while old maids stay on in their mother's house, as 
well ias in the father's or brother's 13 ; ' matara ' 1A was enough 
to designate both parents, and the mother comes before the 
father in such early expressions as ' matara-pitara * and e mata- 
pitarah >15 ; some instances of the precedence of the mother in 
the family 16 are found in later Vedic texts, but they are to ba 
regarded as examples of survival of earlier conditions rather 
than as new developments ; the use of metronymlcs, again f 

1 For references, vide VI., I, 30; II, 113; 486; 495-*96, 

2 Rv. X, 108, 9, 

3 For references, vide V.I.. I. 284-*85. 

4 Rv. X, 117, 9. 

5 Ait. Bra. Ill, 37. 

6 Vide note 1 'above. 

7 Vide note 5 above, 

8 Av. II, 10, 1 (=Taitt. Bra. II, 5, 6, 3) 3 IX, 4, 15; II, 7, 2. 

9 Vide note 4 above. 

10 But * snrm. * is specially associated with c father ' (Rv. I, 1, 9; etc.); 

cf. <the vernacular idiom ; * ma * and * pa,* or * put/ compared 
with ' bap * and * beta ' ; it seems as if * put-ra * (put, po, pola, 
pile) was originally a Bravidian word and hence associated in 
idiom with mother primarily, 

11 Rv. X, 18, 11. 

12 Av. V, 5, 2 ; Taitfc. Bra. in, 1, 1, 4. 

13 Av. I, 14, 2. 

14 Rv. Ill, 33, 3 ; VII, 2, 5. 

15 Rv. IV, 6"* 7; Vaja. -Sam, IX, 19; Taitt. Sam. I 3 3 ? 10, Ij VI, 3, 

11, 3. 
6 Brihad. Upan. IV ? 1, 5 (and in Sutras sometimes},. 

78 y 

though found mainly in later Vedic texts, 1 goes back to the 
period itself. 2 

The practice of polyandry is generally supposed 3 to be 
un- Vedic ; but though absolutely clear instances are not found 
in the Vedic tests, yet certain other customs of Vedic and 
post- Vedic society show evidently po-lyandric traits, so that the 
practice must have existed either side by side 4 or at not a 
very remote age. It has been held that * niyoga s has nothing 
to_ do 5 with polyandry : but it would be niore in accordance 
with natural development to recognize in it a later special 
case^ of an earlier general practice, by which the family 
continuity was assured by all the brothers having an uxor 
communis. 6 The later * niyoga ' is clearly a legal fiction , not 
a new device, but a modification of a wider traditional or 
popular practice : and the custom that formed the basis of this 

* reform ' and theory, must have been a survival of polyandry 
and connected f devr '-marriage. The later ' niyoga/ being a 
restriction, contemplated 'only the begetting of 'a son by the 

* appointed ' kinsman ; but the Vedic l devr '-marriage is not 
so confined : for, in the funeral rite, the son of the widow 
seems to be present, to receive his father's bow, etc., 7 and the 

didhisu 5 Brother-in-law claims her as full wife with no limite'd 
object, but for love, progeny and property generally 8 ; it ia 
obvious that the main concern in this rite is a normal re- 
marriage of the widow, who oftener than n'ot must have had 
borne sons already (the Vedic marriage being one between 
fully developed persons)* ; besides, when the Vedic wife needs 

to^J!? 7 ' v f 6 hn f and ' s absen <* or oft* circumstances, 

to _ continue his line, ,she can have the son through 
other than the devr/'o thoilgh 


\ *** Ij / 4 ,6, i3; of VI 10 2 

6x4t ^ (Indtch.a 3l Lht); they 

^ S m " 

*, a 


-TTT c 

( 79 ) 

SOBS after widowhood by Mm 1 ; so also Mami, preserving and 
following no doubt an earlier tradition, Applies the term 

* didhisu-pati ' to the brother-in-law married to his widowed 
sister-in-law not only for the sake of issue, but also for 
conjugal love, the widow being called ' didhisti ' owing to the 
element of * wooing ' in her second marriage, which is 
recognized as a real one over and -above ( niyoga.' 2 The 
( didhisu ' brother-in-law *s immediate and acknowledged 
claim 3 on the widowed sister-in law, points to the likelihood 
of his having been looked upon as a possible (or even 
secondary) husband before widowhood. Instances of sucH 
view are frequent and clear in the Eplc-Puranic tradition 
referring to the Yedic age 4 ; so also in several passages of the 
marriage hymns the bride is described as * deyr-kama/ 

* desiring union with brothers-in-law '. 5 The epic tradition also 
shows that at the close of the Bgvedic age this preferential 
claim to the widow was not confined to the younger brother 
(as his elder's successor) but also belonged to an elder 
brother. 6 As valay ana's school preserves apparently a much 
earlier Vedic tradition iaccording to which these rights belonged 
not only to brothers, but any other representative of the 
family, e.g. a pupil or a familiar slave. 7 These facts indicate 
that one time several members of a family group, brothers or 
otherwise related, often had a common wife. Thus in ,some 
of ^ the marriage-mantras there is scarcely any sense in the 
bride's being hailed as * virasu devr-kama/ unless the marriage 

1 Rv. X, 18, 8; 30, 2; (the object of attainment of SOBS can only be 

inferred from " janitvam '* in the former passage). 

2 Mann, III, 173. (The simile in Bv. X, 40 , 2, shows clearly the 

* didhisu * character of the widow, and the real marriage she 
contracts. ) 

3 Vido note 1 above; and Av. XVIII. 3, 2-: Taitt Aran, VI, l } 3; 

JLsVal. Grh. Sut. IV, 2, 18. 

** Cf. +He attitude of Brbaspati towards Mamata, and Puskara towards 
Damayanti (vide infra for fuller details). 

5 Rv. X, 85, 44 y Av. XIV, 2, 17, 18, etc.; vid n. 8, p. 80, 

6 E.g. In the case of Bhisma and the widows of Vieitravlrya (vide 


* As'val, Grh. Sut. IV, 2, 18. The inclusion of the * pupil ' shows that 
the custom was specially brahmanic : the famous Svetaketu was he- 
gotten by 'a, * pupil ' (vide infra) j the eligibility of the * dasa * 
for such connections is illustrated in both hrahmanic and nou- 
brahmamc circles, in early as well as subsequent periods : cf. the 
Yv. reference to connexions between Sudras and Arva women ; the 
epic story of the sage Matauga's parentage; the Greek accounts 
about the Nandas, and similar references to pre-Buddhistic court 
scandals in the Midlands in the Jatakas ; and Vats. Ka. Sfit. V, 6, 
12. re. ' dasa' connexions in the foarems, referring to posl-Mauryan 
'and probably enrlier court customs. The commentator on this last, 
distinguishes * dasa ' from * ceta * as * born in the family * and 
' external ' respectively ; the ' dSsa ' of Asval. Gr. Sut, may there- 
fore be such a * dasa/ related to the deceased by blood, and 
hence a good substitute for a brother. (Probably _* jara-dasa * 
would be a better reading than * jarad-dasa * * -or s farai * is to 
be takcto as meaning * hymn-uttering;,* poetic and scholarly, i.e,. 
aa learned as the master ; the sense of * old and senile * would 
be absurd in their contest). 

( 80 ) 

referred to IB those passages is taken to be a polyandrie one , 
where the eldest pf the co-bridegrooms .so addresses the bride, 
alluding to her other secondary husbands, together with 
whom ("we") he hopes to thrive with her 1 ; the "Vedic 
marriage-hymns, obviously do not represent any single standard 
type of marriage 5 but are more correctly a collection of 
mantras of different origins, 2 referring to more than one form 
of marriage, 3 among which the polyandrie is apparently 
Included; so also, in some of the consummation mantras 
14 we " and " men " or " husbands *' in relation to the bride 
may very well refer to these * devr J s * desired by the bride * 
along with the chief bridegroom. 4 In fact the ' sadharani ' 
wife seems to be directly referred to in the Bgveda, 5 where' 
the Maruts are described as enjoying their * common ' and 
eager associate Bodasi, who, with dishevelled tresses .and mind 
devoted to 'her lords, woos them to unite with her, like Surya 
mounting the car of the two Asvins, references to which 
again, are frequent InEv. Specific historical instances indeed 
are not named (as they are in the Epic-Pur anic tradition 6 
regarding Vedic conditions) ; but a few passages 7 probably 
refer to the practice-, specially those where husbands are 
mentioned in relation to a single wife, 8 in most of which 
grammatical or mythological explanations ane inadequate 9 ; 
thus all that is said about the three previous husbands of every 
bride, in the marriage hymns 10 and elsewhere, 11 is best under- 
stood a,s a relic of a gradually disused custom of polyandry, 
which was transformed into an allegory, most probably 

1 Av. XIV, 2, 17.18; 1, 39; Rv. X, 85, 44; of. note 8 below. 

2 Probably often misapplied by the later Sutras; the variant reading 

1 deva-kama ' shows an attempt at conscious emendation. 

3 E.g. polygamy in Av. XIV, 2, 52; vide infra. 

4 Cir. n. 1, p. 11 and Av. XIV, 2, 14.58; Ev. X, 85, 37 (' we * and 

'men 3 ); 38 (patlbhyo jayam). 
a By. I, 167, 4.5. 

5 Vide Infra. 

7 Cl note 5 above; and Uv. VIII, 17, 7 (janiriyabhisamvrtah). In 
Rv. VII, 33, 13, the legend deriving the Vaslsthas and Agastyaa 
from a common mother * Urvasi * and Mitra and Varuna (though 
based on^ an early misunderstanding of names) shows that eminent 
rsi families regarded sharing of a wife by two persons as nothing 
unseemly ; so also, the above reference -to ' sadharani ' wife comes 
from an * Agastya * rsi. "With this may be compared the well- 
known _and nraoh misunucrstood Puranic tradition of the 
Bharadvajas and other brahman gotras being * dvamusyayanaa " 
by origin (vide infra). The biandry in the mitravaruna legend 
has other parallels in the Epics. 

Ev, X s 85, 37.38; Av. XIV, 1, 44.52.61; 2, 14.27, cf. n. 3, p, 81, 
Also Av. II, 36 3 6.7, where a maiden is * given, iiinto husband* * 
(vide infra). 
* 'Hajestia causa*: Weber: Ind. Stud. 5, 191; * generic': Zimmer : 

Alt. Leb. 326 j * mythological ' : Delbruck : Ind. Ver. 543. 
Ev. X, 85, 40,41 =Av. XIV- 2, 3.4. c 

** AY. 17, 2 } for other Bra. But. and quotations, vide Whitney. 
, p. 754. ' * 

.( 81 ). 

re-presenting the life stages of a maiden till marriage 1 ; fathers- 
in-law are mentioned several times in a similar way 2 ; but it is 
uncertain whether polyandry is referred to in any one 
instance,'; there iis however less of uncertainty where at a 
sacrifice 3 the wife is described as ' having noble husbands' ; 
her evident importance and the fertility ritual which includes 
her denuding and wetting in the presence and with the help 
of the conductors of the sacrifice, are probably indications 
that the rite was originally performed by the joint husbands 
of a commpn wife ; so also, a polyandric family custom is 
very flikely referred to in a group- of charms (used to get 
marriageable maidens happily settled), 4 two of which admonish 
the girl to ' turn her right .side to all the responsive suitors/ 
and give her unto husbands.' 5 The striking customs of using 
metronymics (in early as well as later Vedic literature) may 
have originated as much (or even more) in a practice of 
polyandry and laxity among brahman women, 6 with resultant 
uncertainty of paternity, as in that of polygamy 7 ; the former 
view, moreover, is supported by the Rgvedic case of Mama- 
teya the epic case of Draupadeya, 9 and the later Vedic case 
ot Jabala, 10 amongst others. It may be noted in this con- 
nection that the occasional precedence and economic independ- 
ence of the mother seem to be indicated in some Yedic texts. 11 

1 Vide infra. 

2 Rv X, 95, 12; Av. XIV, 2, 27; Katb. Sam. XII, 12. 

3 Taitt, Sam III, 5, 6. ' 

4 Av. II, 36 ; the two passages are w. 6.7. As with the collection of 

marriage-mantras, here too, the charms for securing the marriage 
of girls apparently refer to different marriage customs; thus v, 5 
refers to securing a lover on a ferry-boat (cf . the well-known epic 
case of Satyavati and the Krsn-ite tradition), v. 1 to love-choice 
at ' Samanas % v. 3 to polygamic and v. 4 to monogamic marriages; 
so that vv. 6, 7 may very well refer to polyandric marriage. This 
is followed by * so that she might find one after her wish/ which 
apparent contradiction seems only to mean that the chances are 
that she will find at least one agreeable husband among those to 
whom she is given : a naive defence of the custom surely ! 

5 In Taitt, Sam. VI, 1, 6, 6. it is said that as -women love singers, 

so if there, is a singer in a family, men give their daughters in 
marriage with that family, even if there be others in plenty : this 
however may b interpreted in different ways., though a reference 
to polyandry is possible. The parallel of the Pandava- polyandry, 
where Dranpadi was so given in marriage chiefly on account of 
Arjuna's attainments (musical included) is remarkable. 

6 As shown in priestly as well as non-priestly literature (vide infra). 

7 As supposed by Keith in "Ait. Araij. 244, n. 2. Metronymics may 

also partly have been due to Niyoga (as in the well-known epic 
instances) or to the reputation of women teachers (as in Bra , and 
Upan. sometimes); there may be in the custom a trace of 
matriarchal society, for it is gradually becoming cl-ear that the 
brahrnanic priesthood was originally extra- Aryan (vide infra). 

Rv. I. 147, 3, etc. 

9 Vide Sor. lade*, s.v. 

1& Chand. Uuan. IV, 4, 1.2.4; cf. Sat. Bra. X, 3, 3, 1, etc. 

11 (Vide p. 77, and n. 10-16 tJbere) ; cf . similar indications in the Epic- 
Puranic stories about Bharadvaja and Dlrghatamas (vide infra). 

82 ) 

Widow-burning was practised among many primitive 
Indo-Gerinanic raoes in, Asia, and Europe, 1 a-ndjt can only be 
expected to have e'xisted among the early Indo-Aryans in some 
form or other. But the Vedic literature shows very few 
traces of such a custom : partly no doubt because these texts 
are priestly in character, and widow-burning is known to 
have prevailed elsewhere mainly amongst the non-priestly 
warrior families; and partly because even amongst the ruling 
classes, cases of widow-burning were rare (and prevented) 
throughout the Yedic period, as shown by authentic Ivsatriya 
dynastic traditions. 2 ; while in the s ' brahman J society sex- 
relations seem to have been too lax to admit of the prevalence 
of such a practice. 3 

The hymns of domestic ceremonial and magic in the 
10th Book do not properly belong to the Rgveda, as their 
position amply shows ; they must have been selected and 
abridged from an older and better recension of the Atharvaveda 
or a corresponding collection of traditional ' social * lore (as 
opposed to 'sacrificial'). If, therefore, widow-burning is 
not referred to as an ancient custom in the Ev., whereas it 
is in the corresponding sections of the Av. , 4 it does not prove 
anything beyond this, that the compiler of the former 
to omit certain passages in his abridgment. 5 The full pass- 

?8 'M n ^ f v ** 6 wMcl1 constitut ^s an unit (while the .selection 
IB the Ey.Ms rather abrupt), refers first 8 to the voluntary 
self-immolation of the widow as her ' dharma ' (ancient 
customary duty) , but treats her ' lying down by the departed * 
as only a formal fulfilment of the old custom (though , some 
T< ^ Ve b , een geraine) >- this ' ^B besi <*e ' being 

iS /f Slg * h ,T Pr geny and , 1*!*^ b y something 

legal fiction*; the next verse" makes this attainment 

V .( T ^ ci * n >; *orcopms (De Bello 

_ MM Leb ' 47s * 

^^^^S 11 ^ been supposed (cl VL ^ m ~' g > 

SractiS^fM tni ^ ? 1} tha > eve ? amon gst the Katriyas tho 
S rteSiS U wli aVe ^ 6 ' en univ / sa ^ ^ n g to 4he wastefulnesa 

the S g wifl TM - ^ a ? d the ncess % f *V*g even 

a^aa: vxss; ;?= 

5 ?^I, 3, .1-3- Perhaps also 4. 

* %-v E TTm? ^napwat in the wedding hymn. 

v : Lil > l3 ' 

7 Rv. X, 18, 8. 

8 Av. ZVIII, 3, 

^ T ms o * **> to the 

(videinfrar (m lts prfSent form > in 

'Av, XVIII, 3, 2, 

' '< 88 ) 

of ' progeny and property * possible, by transferring the 
widow as * wife ' to her ' didhisu J who grasps her hand (rais- 
ing and leading her away) ; the ' didhisu ' then expresses 
satisfaction 1 at having saved ' a young woman, enclosed with 
blind darkness, and led about, living, for the dead/ Evidently 
widow-burning was a defunct custom at this time, represented 
only by a ritual ' semblance/ and positively prevented from 
being renewed in any way by an immediate re-marriage. It 
is possible that the expressions in the last passage may refer 
to burning of the widow by relatives, who led her about 
blindfolded ; but this contradicts the first statement regarding 
the widow's own choice ; in any case the rite is deliberately 
prevented; probably only one of these contradictory passages 
was meant to accompany the other (about re-marriage), 
according as 2 the particular case was one of voluntary or in- 
voluntary ' suttee ' ; or, the expressions in question might 
simply be figurative, describing the grief -.stricken and helpless 
state of the young widow. 

It follows from all this that in Vedic society women of 
child-bearing age did not normally remain widows for any 
length of time, being almost immediately re-married 3 ; this 
is probably the force of ' ime avidhavah supatnih ' 4 in the same 
funeral hymn; and it accounts for the rare occurrence of 
5 vidhava ' as such, 5 beside the mention of other widows going 
to be re-married (' gartaruh ') 6 or actually re-married 
C punarbhu '). 7 

The widow often married her brother-in-law and had 
children by him 8 ; this was however not a restricted niyoga ' 
in the later sense, as the widpw's hand is taken formally, not 
only ^ for offspring but also for property 9 ; and she approaches 
the ' devr * as an ordinary young maiden her lover. It is 

* Ibid. 3. The Sutra application of this verse to a cow that is killed 
on the occasion is incredible; v. 4 seems to give social sanction 
to the act in v. 3j ' gopati * is prob. intended as a pun on ' goptr * 
and *pati.' 

2 So also in the marriage hymns, all the mantras apparently do not 

apply fto one type of marriage. 

3 Thus there was nothing unusual in Epic-Puranic cases like that ot 

Ugrayudha (Pancala) wanting to marry the widow of gantanu 
(Kuru), even before the funeral was over (vide infra), for that 
was precisely the- custom. 

4 Av. XVIII, 3, 57 ; Ev. X, 18, 7. 

5 Rv. IV, 18, 12; X, 40, 2; Sadvim. Bra. III. 7; ' vidava.' (like -the 

masc. form ' vidhava/ prolb. Bv. X, 40, 8) would thus seem only 
Jto have designated persons in the 'temporary condition of bereave- 
ment, and not in a permanent state oJ: husbandlessness, 

6 Rv. I, 124, 7 (cT. Nir. Ill, 5.) 

7 Av. IX, 5, 28. 

8 01 n. 1, p. 79. . ! 
I Vide n. 10, p 82, - 

.( 84 ) 

again iiot necessarily a ' devr ' wbo marries her but anyone who 
jw*ht be a k didhisu a ; the widow herself is didhisu indi- 
catin^ some exercise of choice on her part, while her second 
husband is called ' didhisu-pati/ 2 and the .son of such niarnage 
between two pre-eminently ' didhisu (su) ' persons, daidhi- 
savya,' 2 In fact in other references to widow re-marriage 
nothing is said about restriction to the first husband's kin or 
household ; in one of them 3 the previous husband is sought 
to be ignored altogether, and connexion with him cut off in 
the nest world by magical charms, showing 'that the 
1 punarbhu ' is here married into a totally different family ; in 
another, 4 a woman might have several husbands one after 
another, of ' vaisya/ ' rajanya ' or ' brahman castes. 

It appears that apart from regular widow re-marriages, 
women could also re-marry on disappearance of the husband 5 
or in other circumstances in his life-time 6 ; and of the ten 
previous husbands of the widow whom the Atharvavedic 
brahman is willing to marry -as her eleventh and best 
husband, 7 several must have either left her or been discarded 
bv her for various lawful reasons. The number of re-mar- 


riages permissible is nowhere laid down 8 : the custom of 
4 dew '-marriage is no proof for one re-marriage only, for 
similar transferences may well have occurred more than once ; 
the rite to secure reunion in heaven with the present husband 
rather than the previous, 9 if at all believed to be effective, 
would imply similar .safeguarding of every fresh re-marriage ; 
while it is remarkable that in a passage intended to glorify 
the ' brahman/ 10 he should be described as willing to be the 
best husband of a much married widow. 11 

1 (Vide n. 10, p. 82. ) This is taken in the Sutras to include the * devr * 

and other representatives of the husband, like pupil or slave 
(vide ante}. The older Dh. Sutras (often embodying later Vedic 
custom) recognize fully the ordinary widow re-marriage (without 
restriction of sphere). 

2 Of. Maim; III, 173; St. Pet. Diet., s.v 'didhisu/ 3; also 'daidhis.a- 

vya' : Taitt. Sam. III. 2 } 4, 4; Kat. fir. Sut II, 1, 22; Kaul 
Siit. 3, 5; 157, 37. 
5 AT. IX, 5,' 28. 

4 Av. V ? 17, 8.9, 

5 Ev. VI, 49, 8. 

Av. IX, 5, 27.28, may also refer to aach re-marriage (owing to first 
husband being fallen or impotent) : cf. Baudh. Dh. Siit. II, 2, 
3, 27. 

7 Vide a. 4 above. 

8 It is possible, that the allegory of 3 previous husbands of every bride 

reflects, also (vide p. 80, n. 10 and 11, and p. 81> n. 1) 
a contemporary view of the average number (4) of re-marriages 

9 Vide n, 3 above. 

10 Vide n. 4 above. * 

H Which, it is said, was * well-known to the 5 (Manava) races ' ; this 
is quM in agreement with the known facts, Vedic and Epic- 
Puranic ? regarding the character of brahman society (vid infral 

( 85 ) 

Neither of the two different ^iews represented by Zimmer 1 
and Weber, 2 regarding the comparative prevalence of mono- 
gamy and polygamy in the "Vedic age, seems to be a full 
-explanation of the facts,. Thus it cannot be maintained that 
monogamic relations were the normal iand prevalent character- 
istic, for deviations on either side are not rare: e.g., ' sapatni J 
Is found quite early and often^; and apart from indications of 
polyandry 4 and other references to paramours, 5 it is presumed 
by a domestic ritual formula in the Rv. that every married 
woman might have her ' jara/ 6 with which may be com- 
pared similar presumptions in the Yv. and Bra. -ritual 7 ; this 
is also confirmed by the remarkable Epic-Puranic traditions 
regarding Svetaketu and Dirghatamas' reforms, which would 
show that amongst the earlier generations of the Yedic priestly 
society at least, the worn-en were often not ' monogamous.' 8 
So again, polygamy, instead of dying out in the early Yedic 
age, is found all through, and seems to be rather on the in- 
crease, preparing the way for a, greater laxity and corruption 
In the succeeding age. 'Thus ' Maim ' himself is credited 
with ten wives 9 ; Cyavana one of the earliest rsis married a 
number of maidens in old age, 30 and so did Kaksivant the 
Pajiiya 11 in the latter part of the Egvedic age ; while' the Vedic 
prince and his priest who could give and receive scores of slave- 
girls as wives, 12 were no doubt living in an age of flourishing 
polygamy. Cases of polygamy (amongst rsis, princes, or even 
non-Aryan chiefs) are indeed often referred 'to in the Egveda 13 : 
in some of which the relations between the several wives {from 
2 to at least 8) 'and the husband 1 , (' ekah samanah ') are ideally 

1 Zimmer : Alt. Leb. 323. 

2 Weber : Ind. Stud, 5 3 222. 

3 Rv. Ill, 1, 10; 6, 4; cf, I, 105, 8; X, 145, I. 2. 5; (besides 

Av. frequently). 
4, Vide ante. 

5 Vide infra. 

6 Rv. X, 162, 5. 6 = Av. VIII, 6, 7. 8. 

7 The wife's questioning and confession: Mait. Sam. I, 10, 11; Taitt. 

Sam. I 3 8, 3 ; Tartt. Bra. I, 6, 5, 2 ; gat. Bra. II, 5, 2 S 20. 

8 Vide n. 1 3 p. 74. 

9 Mait. Sam. I, 5, 8. (Vide infra for Puranic notices of the polygamy 

of Mann and his descendants). 

10 Rv. I, 116, 10; (with gat. Bra. IV, 1, 5, 1 ; 10, 13; Jaim. Bra. 

Ill, 121 fi) ; cf . V, 74, 5 ; and allusions to above in I, 117, 13 ; 
118, 6; VII, 68, 6j 71, 5; X, 39, 4. (These ' kanfs and 
* vadhu's were over and above the famous princess Sukanya). 

11 Rv. I, 126, 3 (10 ' vadhumant ' cars f rom Svanaya) ; cf . I, 51, 13 

(Vrcaya in old age). 

12 Rv. VIII, 19, 36. (Trasadasyu-Paunikutsa and Sobharl-Kanva may 

belong to the earlier part of the Rgvedic age; but Pargiter 
places them in the latter part, distinguishing 2 Purukutsas and 
2 Tr&sadasvusl 

IS Rv. I, 62, 11; 71/1; 104, 3; 105, 8; 112, 19; 186, 7; VII, 18, 2; 
26, 3;X,M3, 1; 101, 11; {It is remarkable that almost all these- 
references to polygamy come from Angirasa and Vasistha rsis). 
Of. other references in Av, and Yv. : e.g. Av. III ? 4j etc., 
Taitt. Sam. VI 3 5 ; 1. 4; etc- 

86 y 

happy, while in others they are recognized as painful. the 
circumstances of conquest and settlement, and consequent 
prosperity of the priesthood, must have made polygamy ;a 
common thing. It is significant that in the Bv. dasa is 
primarily the enemy and only secondarily a slave, but that 
' dasl ' is all along the ' slave-girl ' from the Av. onwards 2 ; 
this would show that the first slaves were the captured Dasa 
women, slave-concubinage developing quite early side by side 
with the Aryan conquest. 3 In the later Samhitas the slave 
woman is also called ' sudra ' 4 (probably originally a term 
of racial significance like '_dasi T and such a ' sudra ' often 
rose in the favour of her Aryan master 6 who must have had 
his Aryan wife or wives. 7 The earlier Brahmanas _ directly 
ascribe ' sudra ' or ' dasl ' concubinage to eminent ' rsi ' fami- 
lies (Egvedic as well as more or less contemporary ones) , and a 
' dasl- (or sfidiaO putra, 1 though subject to natural com- 
ments, was nevertheless common enough to be assigned the 
same position as other rsis and teachers. 8 In the Kgvedic 
texts themselves, female .slaves are frequently presented to 
pgiis by their patron princes ; thus King Trasadasyu 9 bestovred 
fifty of them as ' vadhu 's on Sobhari-Kanva 10 ; and in other 
cases, presents of horses, camels or buffaloes, are embellished 
by such * vadhii ' slaves along with them 11 ; while chariots are 
'described as full of slave-girls. 12 The number of slave-girls 
kept in single establishments in no way diminished in the 
next age : thus the Satepatha knows of as many a.s four hundred 
* anucari ',s 13 ; and (even) in the Aianyakas and Upanisads 

1 Bv. VII, 86, 7; VIII, 56, 3; X, 62, 10; prob. I, 92, 8; 158, 5; 

VHI, 46, 32; cf. Av. IV, 9, 8. 

2 Av. V, 22; 6; XII, 3, 13; 4, 9; Cband. Upan. V 3 13, 2; Brhad, 

Upan. VI, 1, 10. 

3 Which was probably over before the later Sambitas and Bra ., for 

they do not refer to any Arya-Dasa wars, but only to Aryan 
wars (vide V.I. ? i, 65). 

4 Av. V, 22, 7; Taltt. Sam. VII, 4, 19, 3; Kath. Sam. (Asvamedha), 

IV, 8; Mait. Sam. Ill, 13, 1; Vaja. Sam. XXIII, 30; etc. 

5 Vide V.I., II, 392. 

6 Vide Yv, references in /iote 4 above. 

7 Who also bad connexions with Sudra slaves ; Vaja. Sam. XXIII, 31. 
a E.,g. Kaksivaatj sou of a slave-girl: Brfaadd. IV, il-15; 21-25; 

with Bv. I, 18, 1; 112, 11; 140-164 (cf. Pane. Bra. XIV, 11, 
I6J ; Kavasa, 4 dasyah putra^t ' : Ait. Bra. II, 19, 1 ; Kau. 
Bra. XII, 1,3; (Kavasa was a Rgvedic i'i) ; Vatsa, ' sudra- 
putra J (a Kanva) : Pane. Bra. XIV, 6, 6 ; cf . Satyakama Jabala 
(about 150 years after the compilation of the Kv.) : Chand. 
Upan, IVj 4, 1 if ., etc. ; also in Brbad. Upan., Ait. Bra, 
and Sat, Bra. (vide details of ref. in V.'l., H 5 420). 
9 Bv. VIII, 19, 56. Cf. V, 47, 6 (' vadhu's). 

10 Sons by slave-concubines was a special feature in the Kanva 

groups ; vide V.I,, II, 238. 

11 Bv. VHI, 68, 17 ; VI, 27, 8 ; Av. XX, 127, 2, * 

12 Bv. I, 126, 3; VII, 18, 23. (These * vadhu's however might b* 

* slave * aa well as free, from the context). 
H Sat. Bra, XIII, 5 a 4, 27. 

{ 87 J] 

the King is attended by five hundred fair women carrying 
perfumed powders^ etc. 1 The presence, increase and distri- 
bution of slave women was thus a fertile source of polygamy 
among princes and priests alike. 2 

Apart from this possession of slave-girls, the princes had 
at least 3 four principal wives recognized in regal ceremonial 
and rites, of whom the fourth, the k palagali/ seems to be 
a comparatively later development, or to have been given a 
place in the ritual somewhat later in the Brahmana age ; 
the ' mahisl ' and the ' parivrktl ' occur from the [Rgveda 
onwards 4 ; and though the 4 vavata ' first occur.s in the Av., 5 
she is implied by the ' parivrktl ' ; the * palagali,' wife of the 
King is an indication that it was. a- political marriage, 7 and that 
daughters of other and higher court officials also were custom- 
arily taken into his harem from similar original motives 8 ; the 
first three designations are essentially relative, and pro-suppose 
a regular harem-establishment, the members of "which experi- 
enced constant rises and falls ( ' parivrkti ' ) in power at court 

1 Kau$. Upan. 1, 4; and corr. passage in Sankh. Aran. 

2 Vide Puranic parallels infra. 

3 Tihe King's many wives are referred to in Rv. VII, 18, 2 (' rajeva 

hi jauibnia'}; probably .' konibhik ' in iiv. X, 9o 5 9, icters to 
Pururavas ' other wives besides Urvasi ; of. Rv. II, 16, 3 
(Indra's * indriya ' not overcome by his ' ksoms *). 

4 ' Mahisi ' : Rv. V, 2, 2 5 37, 3 ; Av. II, 36, 3 j Taitt. Sam. I, 8, 

9/1; Mait. Sam. II, 6, 5; Kath. Sam. XV, 4; Tai'tt. Bra. Ill, 
9, 4, 4; Pane, Bra. XIX, 1, 4; Sat. Bra. V, 3, 1, 4; VI, 5, 3, 

1, 6; XII, 2, 6, 4; 1, 8; 5, 2, 2. 5. 9; k panvrkti/ etc. : Rv. 
X, 102, 11; Av. VII., 113, 2 ; XX, 128, 10. 11; Karh. &am- 
X, 10; XV, 4; Taitt, Sam. I, 8, 9, 1; Taitt. Bra. I, 7, 3, 4; 
Sat. Bra. V, 3, 1, 13; XIII, 2, 6, 6; 4, 1, 8; 5, 2, 7. 

5 Av. XX, 128, 10.11; subsequently in Taitt. Bra. I, 7, 3, 3; 111, 

9, 4, 4; Ait. Bra. HI, 22; gat. Bra. XIII, 2, 6, 5; 4, 18; 5, 

2, 6 ; in fact -the Yv. ceremonial presupposes -these 4 chief queens. 

6 Pailagall: Taitt. Bra. I, 7, 3, 3 ffj III, 9, 4, 5; Sat, Bra. XIII, 

4, 1, 8; Sankh. gr. Sut. XVI, 4, 4 

T She is the daughter of the lowest court official, probably the chief 
'palagala' (messenger or spy) (Sat. Bra. V, 3, 1, 11), whosj 
function is aptly described as bearing false news ; the motive of 
takin such an officer's daughter in the harem is quite clear ; 
and as the ' lowest ' officer's daughter is a queen, other officers 
daughters also must have been favoured, as indeed is evident 
from the numerous companions of the 4 chief wives (present at t^e 
horse-sacrifice) belonging to different ranks. f. A^vamedna 
sections of Yv. Samhitas. , 

ft Vide n 7 above. This is illustrated in Epic tradition also; thus 
Sumitra the * parivrkt!' wife of Dasara/tha was the 'ptufohita, 
Vamadeva's daughter by a ' vaisya '), and one of (the Matsya 
kine^ Virata's queens was a sister of his commander-in-chief 
Kicaka. In" later literature- ' Mahamatra-suta's are often taken into 
the royal harems (cf. Vats. Ka. Sut.). The ' vaisya wife 'so often 
mentioned in Epic-Puranic tradition, is probably th_e daughter 
S the ? Gramanr of the king's court, white th e ' .^ta ^ifo 
is the daughter of the Palagala or lowest court official, the spy- 

( 88 ) 

( f mahisi ' ) or in personal favour with the prince ( ' vavata ' ) - 1 
Such rise and fall is well depicted in the chief wife's song of 
triumph/ where she congratulates herself on the dawn of her 
fortunes, subjugation ot rival wives and influence over the 
heroic lord with whom her name stands highest, and through 
whom she rules ell the people, on her sons rising to the 
rank of mighty warriors and daughters to that of princesses. 

This threefold classification seems to have been a general 
one, and not confined to consorts of princes : thus a domestic 
mantra wishes tliat a maiden might after marriage become -a 
mother of sons, and thereby become a ; mahisi ' a ; while the 
rivalry between the ' vavata ' and ' parivrkti ' wives forms 
the subject of many other domestic magical rites. 4 Three- 
wives then would appear to have been a common average, 
almost a minimum lor the Vedic polygamist householder, 
though two wives are mentioned once in the Kv. 5 ; so also, 
in some early Brahmanas, mention is made of the sons of 
one's father's eldest wife and youngest wife (' jaisthineya J 
and ' kanisthineya 'j. 6 In a passage of the marriage hymns 
several young maidens are .said to be eagerly proceeding to a 
husband's home from their father's (or fathers'), 7 wherei the 
reference evidently is to one man marrying several sisters or 
otherwise related women .at the same time, With the Yajur- 
vedic brahman indeed, ' " many wives ' ' was an apparently 
established custom, 8 Of a man's several wives one at least 
must often have been the widow of a brother or kinsman, 
from the customary character of such transference. 9 Apart 
from these, regular wives, the example of slave-concubinage 
amongst princes and their client priests 10 must have influenced 
ordinary society; the references to Arya-Sudra unions in the 
Iv. bamhitas is rather general, and might imply that slave- 
women -were glad to be wives of any Arya whether rich or 
poor (tor obvious advantages) ; the employment of ' dasl 'a or 

1 Here also the Epics afford interesting illustration, e.g. in the cliamr- 

fflg re at-ions between Dateatha's 3 wives and Krsna's many 
jvives (at least two of whom were maiiisi > by tons, and' 5 

\ JtV dlt3i J , 

2 RT. X, 159. 

3 Av. II, 36, 3. 

4 AT, III, 18; VII, 35; Rv. X, 145. 

5 Bv, X, 101, 11. 

7 If XW% n 52 ls 8> lj PaliC " Br ' Hj l > 2j XX > 5 > 2 - 
' ' vahvir i5y ^ 

ID Vide ante. r 

II For when t a flud* became the beloved of her Arya lord, she did not 

( 89 ) 

8 siidra 's as * aimcaii 's and ' parivestrl 's 1 must have become 
a common item of style"; even the ordinary brahman 
sacrifice!, while placing five conical bricks 2 on his fire-altar, 
hoped to obtain in the next world five fair ' asparas'es as his 
personal attendants, bodyguards and k embracers/ 3 evidently 
the heavenly counterparts of his humbler .establishment ; it is 
also probable that already in the Vedic marriage the 
' uyocam ' 4 refers to a companion slave-girl given away along 
with the bride, a very ancient custom ascribed to some of the 
earliest royal marriages in. Epic-Puranic tradition. 5 

While however ' a general prevalence of monogamy ' or 
the * dying out of polygamy ' are not borne out by such facts 
and indications, it is reasonable- to hold that as polygamy- 
must always, in the absence of universal regulations, be 
secondary with communities, tending to appear pr disappear 
according to variation of circumstances, it has had this history 
in (ancient India. Thus it may well have existed in the 
primitive tribal stage of the Aryans, when large numbers of 
women of subordinated kindred or enemy groups may have 
been transferred to mighty horde-leaders or patriarchs 6 ; it 
would develop with the extermination or assimilation of the 
Dasas 7 of the plains in the early Vedic period; it would 
become a fashion subsequently with the growth of an opulent 
ruling nobility and their favoured priesthood, 3 or be inevitable 
with the progress of internecine fighting 9 ; it may have been 
adopted in the earliest times from pre-existing non-Aryan 
princes and priests. 10 But between these secondary develop- 
ments of the Vedic age a monogamistic tendency seems to have 

1 Vide 11. 13, p. 86; Sat. Bra. XI, 2, 7, 4; Kaus. Upan. II, 1; Keith : 

ankh. Aran., 21, n. 2. 

2 A curious parallel to this association of ideas is to be found in t-he 

*' baa "-wives of ' baby ' in Bengali household idiom. 

3 Taitt, Sam. V, 3, 7, etc. , _ 

4 Av. XIV, 1, 7=Bv. X, 85, 6; being classed with / anndeyi/ jit 

must mean companion-maid (represented by the ' jhi * or^ * dasi 
of even modern times), rather than any ornament or special type 
of song, t 

5 E e Sarmisthi, and her maids given away to Yayati along mto 

Devayani; or similar gifts in the case of Draupadi and Suba- 
adra's marriages. Vide infra. _ . 

6 The case of Manu's 10 wives would fall unSer 'this head; for Puramc 

instances of the polygamy of such early chiefs (like Paksa, 
Ka^yapa, Maiiu, Iksvaku, etc.) vide infra. 

7 Of. pp. 8587 above. 

I This P ^s M^iSated by Epic-Puranic instances : vide into. 
10 This probability becomes almost a certainty when, some of the tradi- 
tional ' instances of polygamy are critically wwed : vide inf la. 

( 90 )] 

been always present, 1 and the persistence of this ideal is 
discernible'' through all the fluctuations of subsequent periods. 
In this matter indeed, ancient Indian society has developed 
and changed unfettered by any external cpnimandment or 
ruling (unlike society in Europe and the Middle East, where 
a monogamic and a polygamic character, respectively, has 
practically been imposed * by Christianity and Islam) ; and 
prevalence of polygamy or monogamy for any particular period 
or region has depended on various communal, economic and 
political conditions, and the state of public opinion or 
individual ideals. Thus it is intelligible how side by side with 
instances of polygamy and laxity, monogamy is evidently 
approved in the Egveda as an ideal 2 ; constancy of conjugal 
affections is earnestly sought for equally by men and women 3 ; 
while a large portion of the wedding-hymns (scarcely surpassed 
by any other nuptial formulas for simple yet noble ideas) 
regard the marriage-tie with reverence, and, practically 
ignoring polygamy, emphasise mutual conjugal fidelity, 
poetically typified 4 in the * cakravaka ' pair. 

1 Cf. the use of _' patni ' in the singular; and -the recognition of only 

one full wife in litual (patni) or at royal court (mahisi). 

2 Ev. I, 124, 7 3 IV, 3, 2 ; X, 71, 4 ; etc. (apart from the marriage 

5 E.g. AY. II, 30, 2. 5; 36, 4; VI, 139; VII, 36; 37; 38; cf. VI, 

102 ; 130; 131; 132; (apart from the marriage hymns). 
4 Av. XIV, 2, 64. * ) > 




The Vedic marriage Is a natural and a real one, with 
little of the rigidity .and artificiality of the later ' Hindu * 
forms. The only possible (?) reference to an early marriage 
is in an Upanisad, "where a poor brahman teacher adopts the 
life of a beggar with his ' atiki * wife 1 : the medieval 
commentators give ' atiki ' a fanciful special sense, of * ajata- 
payodhara, etc./ which evidently reflects their own dislike 2 
of the idea that a brahman teacher's youthful wife should go 
about freely ; if it is not a proper name, and has to be taken 
11 s an adjective, the only rational sense would be ' fit for or 
used to a wandering life/ i.e., hardy and patient 3 . Qhild- 
wives are first mentioned in the Sutras 4 ; and there the 
gradual growth of the practice may be clearly traced, from its 
beginnings in the time of Asval. vand Hiran. Sutras onwards; 
even then child-marriage had not become a general rule. 5 This 
1 legal ' Sutra evidence is borne out by the ^post-Mauryan) 
Vats, 'Kama'-Sutra, which ignores child-marriages altogether, 
recognizing in special cases juvenile attachments and wooings 
only. 6 It seems probable that this subsequent cropping up 
and development of child-marriage as a practice was due to a 
certain amount of insecurity of society 7 in the earlier and latter 
parts of the ' Sutra period/ between eir. 550 and 320 B.C., 
and from 220 B.C. onwards, as a result of Persian and 
Macedonian conquests, and Graeco-Bactrian, Parthian, Scythic 
and Kusan invasions, respectively/ 

In the earlier Vedic period, the obligatory marriage of a 
girl, before a certain age, and irrespective of all other con- 
siderations, was unknown. 9 Thus, forward younger sisters 

I Ghand. Upan, I, 10, 1. 

3 Acquired in dissimilar social and political circumstances. 

3 The S.B.E. however, adopts the view of the commentators. It seems 
permissible to see in ' atiki * a reference to ' itinerant ' women 
teachers , (married or otherwise: vide infra.), who are also well 
known to Epic-Puranic (tradition. Of, the Yedic ' Itant ' or 
' Ita * rsis and the * yayavara's. 

Of. Jolly : Itecht nnd Sitte ; 59; Hopkins ; J. Am. Or. S.. 13, 340 ff; 
23, 356. 

5 Bhan^arkar: Z.D.M.O. 47 ; 143-156 (in review of Jolly : ibid. 46; 


6 Cf. specially, Vats. Ka. Sut. Ill, re "wooing of the ' kanya V* 

7 It is well-known that early marriage became general in medieval 

India largely owing 4o the Mahomedan occupation of the- country. 

8 Subsequently, however, child-marriage must have fallen into disuse, 

specially during the Gupta period (aa the^ evidence o^ Gupta 
literature generally shows); It would revive- again with the 
collapse of Indian polity before the invasions of the 6th and 7th 
centuries; and before a full restoration of normal forms, th& 
special feature would be confirmed by Mahomedan invasions 
and subsequent occupation. 

9 Except possibly in the case of royal alliances, where occasional _early 
marriages may have taken place, naturally enough. Vide infra, 
re indications of it in ( tradition '. 

( 92 ) 

might get married in advance while the eider still waited for 
her chances in love 1 ; cases ol unmarried young women stay- 
ing on with their father, and even growing old (or dying 
unmarried) 2 in the paternal home, were not unusual, 3 though 
an old maid was regarded as rather unfortunate, eliciting 
ironical remarks (e.g., being called Yama's ' Eulapa/ 4 or 
' sitting long with the Fathers J ) , and maidens cursed their 
rivals in love with hated spinsterhood. 5 

The early Vedic texts 7 know of mutual affection, develop- 
ing between the youth and the maid. Thus, the love-led 
maiden (jarini) goes to her tryst, with as strong a passion 
as that of the gambler for his dice 8 ; the river offers an easy 
ford, as a ' kanya ' bends herself to receive her ' marya's 3 
embrace 9 ; the young woman wcos and attends her dear lover/ 
and the fingers press the ' Soma ' as a ' kanya * caresses her 
lover. 11 Young people dream of the co-mingling of body, 
intents and conduct, of the woman desiring a, husband and the 
man desiring a wife coming together in joy and blessedness 12 ; 
parents wish that their marriageable girl may find a husband 
according to her wish and choice and responsive to her love, 15 
and at^ the same time be enjoyed by, clear to, and -concordant 
with him 14 ; and with couples about to be married, the eyes of 
both_are of honey-aspect/ 5 their faces ointment, they are put 
within one smothers heart, ,and their minds are together. 16 

On either side the yearning described is that of persons 
IE the fulness of youth. Thus, the sun follows the dawn 
like a youthful lover after an attractive woman 17 ; India is 
coaxed as a confident lover proud of his * yosa ' coaxes her 18 
the youth imagines his chosen girl as pierced with Kama's 
shafts (feathered with longing, tipped with love, necked with 

1 Vide infra. 

2 This was not a dreaded fate In early Vedic estimation ; cf. Av, 

A\ 111, 2, 4-7. 

3 Bv. I, 117, 7; II, 17, 7; X, 39 ; 3; 40, 5; Av. I. 14. 

* I.e. _ mistress 01 Pluto's household ' ; similar remarks are still in uas 
IE rern. idioms. 

5 In a double sense, 
g Av, I, 14, 3. 

7 In the following lines the original texts have simply been paraphrased. 

JSv. 2L ? 0-3, 5; cr. 40, 6. 

9 Bv. Ill, 33, 10, 

10 Rv. IS, 3?, 5. 

11 Bv. IX, 56, 3. 

12 Av. II, 30 3 2-3. ('spouse-finder 1 mantras). 
IS An oft-repeated phrase 

14 Av. II, 36, 4.5 (etc.) ; cf. VI, 60, 3. 

15 Cf. Av. I, 34 (emphasising * sweet ' relations). 

tt AT,, VII, 36; cf. VI, 102 (moving together like, a king-horse and 

a side-mare). 
1? Ev. 3 I, 115, 2. 
18 Hv, s TV, 90 5, 

( 93 ) 

resolve, consuming, humbling, etc.), so that impelled away 
from her pa-rents, and leaving her cosy conch, she comes to 
him creeping, gentle and sweet, and entirely his 1 ; he wants 
her to burn and dry up with desire for every limb of his, lust 
after him, and cling to his arm and heart 2 ; on the other hand, 
the maiden also wants her man to think of, pine for, and be 
mad after her, while she would not fall in such plight her- 
self, though he is dear to her 3 ; in fact the young man often 
loses his head and makes a present of all his belongings to 
his girl 4 ; she too, believing that it is after the manner of the 
gods themselves, and in accordance with Varuna's ' dharma/ 
baldly kindles the flame of burning love. 5 

On either side, again, strong jealousy is felt In love- 
affairs, and Wia-ndering affections are anxiously sought to be 
recalled, which shows much freedom of intercourse. Thus 
rival maidens cursed one another ceremonially with .spinster- 
hood, 6 and malicious rites were performed by men also 
against their rivals 7 ; when going abroad, the young man is 
reminded 1 by his sweetheart that he is wholly hers, must 
never even mention any other woman, and must return to her 
even from beyond unknown lands and streams, and he must 
not say 'anything against this prayer of hers, for a man's talk 
stilts only the assembly, but he is to be quiet before Ma 
sweetheart 8 ; when the lover ha actually left her, she still 
wants him to long for her with his whole body, come back 
to her and be the father of her sons, though he may have run 
5 leagues away, or a horseman's day's journey 9 ; and the 
jealousy of rivals in love is reflected in the rite where tte bride 
symbolically binds her groom with her hair to make him 
wholly hers, so that he may not henceforward even name an- 

fel J 

other woman 10 ; on the other hand when the maiden proved 
inconstant, her jilted lover earnestly hoped that she might 
yet dry up in heart and mouth by loving him, and that 

1 Av. Ill, 25. 

2 Av. VI, 9; cf. VI, 139 and VI, 8. 

3 Av. VI, 130. 

4 Ev, I, 117, IB. 

5 Av. VI, 132. 

6 Av. I, 14. Tlie rite for barrenness of a rival woman might also have 

been performed by such jealous maidens; cf. Av. VII, 35; also 
VII, 113 (mutilation). 

7 Cf. Av. VI, 138; VII, 90 (inducing impotence; performed also 

against wife's paramour). 

8 Av. VII, 38 (might also be used by wives) ; Whitney refers to 

" Burmese " parallels of -the 'thread-tie ' ; but cf. the 
' rakhi* throughout the " Gangetic " country. 

9 Av. VI, 131. , 
19 Av. VII, 37, 

( 94 ) 

estranged hearts might nevertheless be joined together and 

made the same, 1 

4part from these plain descriptions, the very fact that 
there were regular domestic rites (with charms and magic 
potions) 2 calculated to help in all the momentous stages of 
the progress of love-affairs, 3 and that e\en the guardians of 
maidens took part in some of them, 4 shows that free love- 
makings between young men and women before marriage, 
was fully recognized in ordinary society. 

Good opportunities were afforded for these pre-marital 
loves in the Vedic festivals. The ritual of the Mahavrata 5 
shows that it was the Bralim'anical counterpart of some 
popular spring festival, 6 wherein there was much of song and 
dance, swinging and free intermingling of men and women, 
running into the extremes of promiscuity. 7 But apart from 
such orgies, there was the more decent 8 group of mixed gather- 
ings called ' Samaria's, 9 where the most prominent feature was 
the wooings of lovers with a view to matrimony, 10 and the 
lighter pleasures of the company of the fair sex in their most 
agreeable mood and choicest attire, 4 though events like 

1 Av. VI, 159. 

2 Some of these rites have "been indicated in the above references. 

3 Viz., acquaintance > growth of lov^ secret visits, jealousy, estrange 

ment, reunion, etc, 

4 E.g. Av. II, 36 ; VI, 60. 

5 Tide the * Mahavrata * sections in Yv. Samhitas. 

6 Ci. Keith : Sankh. Aran., re the Mahavrata. 

7 This may have been the prototype of the classical Hallisaka and 

Latarasaka, mentioned in Vats. ICa. Sut. as specially suited for 
courtships in polite circles, and of the more vulgar ( ?) medieval 
and modern Holi (Holaka, Dol, etc.), and Rasar (Jhulan, etc.). 
But the inclusion of martial features in the Mahavrata, and the 
sort of drum and dancing described, rather point to some Dravido- 
Kplarian affinities*, cl the seasonal orgies of the N.-E. Deccan 
tribes; the extremes of licentiousness (bhutanam maithunam) are 
common .to these as well as to the other group of festivals.' 

8 Not always, for the sessions sometimes lasted the whole night, and 

girls spent the night out there; besides courtesans also tools 
advantage of these * Samana's. (Rv. I, 124, 8; cf. 126, 5j 
brilliant e vra's attending the Samana; { visyah vrah ' with many 

9 Probably the e Samana s was primarily a seasonal festival, at the 

beginning of the * sama ' or summer, which came to serve as 
the occasion for various social functions (just as even now mar- 
riages mostly take place in the months wherein the Vasanta- 
"Paucami and Holi fall). Indrani (a sex goddess) was worshipped 
bv women at these Samanas according to anoieut custom : Rv. X, 
86, 10. B 

10 Av. EL 56, 1 (agreeable and enjoyable to suitors) f Bv. VII, 2, 5 

(adorned all over); Ev. IV, 58 5 8 (amorous, smiling, auspicious, 

eie,)j etc. 

( 95 ) 

poetic contests, 1 tournaments, 2 horse-races, 5 or weddings, 4 may 
have -served as occasions for the gatherings. 5 The fire-lit 
night 6 of such ' Sama-na's witnessed, among many other gay 
and knightly scenes, 7 those of young women (' kumari's 
enjoyable to suitors) making love, 8 and heavily adorned old 
maids (' agravah ') seriously in quest of a husband. 9 Among 
these young and elderly women must have been the * agre- 
didhisu/ or the younger sister who anxious to marry would 
not wait 10 for her elder sister, the ' didhisu/ the le,ss lucky 
elder sister knowing better rather late, 11 or the widow wooing 
afresh ' maryam na yosa/ 12 as well as the spinster growing 
old at home and staying with her father, 13 and the forward 
brotherless girl 14 

Such social freedom is characteristic of the early Vedic 
period, seclusion of women being unknown ; even after mar- 
riage, wives, who ordinarily move about well-adorned within 

1 Possibly dramatic dialogues (sampled in the Bv.) were also acted in 

these Samanas; a piece like Pururavas-UrvasI or Yama-Yamj 
would be particularly suited for such audience (the later 
' Yatra * ' Kavi-gan,' etc., represents the Samana in this aspect). 

2 Such contests were probably followed or occasioned by * svayamvaras % 

as frequently in the Epics and Puranas ; cf . * samanartisu ' in 
Av. XIV, 2, 59 if, where mock fighting (for the bride) at th 
Samana or marriage assembly is referred to (vide infra). In the 
Epic pure tournaments also are attended by ladies who have seats 
in high galleries. 

3 Of. Kv. X, 168, 2, where mares at the Samana run with the Wind 

who rides on them like an universal king. 

4 Av. VI, 60, 2 (cf. XIV ? 2, 59 t.) (' Samana ' here may mean a 

' svayamvara ' assembly as well) ; maidens * toiled to attend 
these ' to help their own causa. . 

5 The Samanas have been compared to Greek festivals; ancient festivals 

are naturally more or less similar; but the parallel of Dravido- 
Kolarian festivals is at. least equally striking, and ' nearer home. 
There is in them the same martial elements, free love-makings 
and excesses, marriages by capture and mock-fights, all-night 
revelries, and a remarkable passion far attending them in choicest 
attire, with young men and worffen alike; all the 'Samana 
imagery * in Vedic literature can be applied equally to a festival 
like the 'Kol-yatra'. The Greek festivals also were based on 
earlier non-Indo-European institutions. 

6 Rv. L 48, 6 (Dawn dispersing the Samana) ; VII, 9, 4 (fire blazing 

bright .at the Samana like the sun) ; cf. X, 69, 11. 

7 Cf. Pischel : Ved. Stud. II, 314. 

8 Av. II, 36, 1. 

t^^K'lt^- 8: "dJW 40 ( 

' Bvayamvaram "I ; Kulluka on Mann. Ill, l<6O.and comm. on Apast. 

la Cf. Ev. t No 1 ' I': ^A in Manu. Ill, 173 ; vide St. Pet. Diet., 
s. v. " didhisu', 3. 

.. 4. 7; Av, I, 17, 1 (in red garments); of. Bv. IV, 5, 5- 

{ 96 ) 

the house, often came out to the Sabha. 1 The maidens grow- 
ing up in their father's home mixed freely with the youth of 
the village, 2 and with them joined in the rustic music and 
swings under the spreading banyan-trees 3 ; the virile young 
man (' marya ') is normally a lover, constantly in the company 
of youthful maidens (' yuvati/ etc.), 4 and, like the latter, 
affects bright and attractive costumes 5 to enhance his ' marya- 
sri ' ( f lover's grace ') 6 ; on the other hand, the young maiden 
is also fully engaged in the midst of a number of suitors, 7 
trying her best to please and attract them at the Samana, 7 
on the ferry-boat/ or at home, 7 turning her right side to 
every responsive suitor 7 ; she meets her chosen lover at trysts, 5 
and lies only half asleep -at night, expecting him to come and 
awaken, her 9 ; the bold youth also secretly visits his lady-love 
in her own chamber late in the night, while all her kinsmen 
are asleep, remaining with her till dawn, 9 

Thus it is only to be expected that the early marriage 
ritual also presupposes that the married pair are grown up 
enough 10 to be lovers, roan and wife, and parents of children, 
and to begin a full home life of their own 11 ; almost at every 
step of that ritual, formulae are repeated showing thear im- 
mediate fitness for procreation 12 ; and ' handgrasping ' and 
consummation are the essential parts of the Vedic marriage. 12 

1 Bv. I, 167. 3. 

2 V.I., II, 485. 

3 Av. IV, 37 ,, 3-5; the green and white swings (i.e., festooned with 

leaves and flowers), the music of cymbals and lutes, or the crests 
of peacock-plumes, ascribed to Gandharvas and Aparases, under 
the Asvattha and Nyagrodha. can only be & reflection of ordinary 
village merry-makings. 

4 Kv. Ill, 31, 7 ; 33, 10 (embracing ' kanya ') ; IV, 20, 5 (flattering 

* yosa ') ; IX, 96, 20 j etc. 

5 Rv. IX, 96, 20. 

6 Kv. II, 10, 5 (in enviable colours). 

I Av. II, 36; etc. 

8 Bv. X, 34, 5; etc. 

9 Bv. I, 134, 3. 

* Uninjured and unexhausted ' : Rev. VIII, 55, 5-8= Av. IV, 5. This 
forma! rite would show that such clandestine meetings were 
common and connived at in society. 

& Cf. * pati-vedanau * (' spouse-finders ' = the breasts): Av. VIII, 6, 1. 
Sonie^ Grh. Sutras (acquainted with the later practice of child- 
marriage) plainly declare the Vecfic marriage-ritual -to be unsuit- 
able, being meant for adults only; but even for that period, 
cf. the essential qualification of the bride in Vats. Ki. Sut., 
s stant* 

II Ev. X, 85 1 Av. XIV, 1 and 2, It is not improbable that several 

passages in^ the marriage mantras (Av. XIV, 2. r 22-24) really refer 
to a legalising marriage after the -woman; has borne a son, who 
also Is thus given the rights of primogeniture, 

( 97 ) 

Fully In accord with 1 these features, there is little trace 
of any real parental control 1 over such mature marriages, 
The later custom of parental sanction would become a necessity 
only as child-marriage became frequent. It is however prob* 
able, from the cases of Syavasva and Yimada, that such con- 
trol where it existed was more a characteristic of the Vedic 
ruling nobility (for obvious reasons) than of the Vedic priest- 
hood, 2 which seems to have been generally indifferent to such 
eugenic considerations. Parents had to submit themselves to 
their new daughter-in-law's rule, she becoming at once the c em- 
press ' of the household 3 : this position she could hardly have 
attained if the son was normally married at the dictation of 
his parents to a ' given away ' girl. There is no evidence that 
the son's marriage could be legally controlled by the father, 
and not much of it in the case of the daughter. 4 But parents 
often had a share in arranging suitable matches, as Arcana- 
nas had, 5 acting as a f vara ? or intermediary 6 in the wooing 
of his son Syavasva who could not hope to succeed all by him- 
self ; so also the mother seems to have had a share (amount- 
ing sometimes to control) in the selection of a husband for 
her daughter, whom she helped in her toilet to make her 
acceptable 7 to suitors; Eathaviti Dalbhya's queen objected 
to her daughter's marriage with Syavasva (though the Mng 
was quite willing) insisting that her son-in-law must be a poet, 
so that the rejected candidate had to become one 8 ; the gambler 
in the Egveda counts it a great misfortune to have lost the 
favour of his mother-in-law, 9 which may have given him his 
wife. But sometimes an ardent but otherwise undesirable 
suitor (' vijamatr ') had to please the father by heavy pay- 
ments for his bride 10 ; or conversely, if a maiden had any 
defect or was unable to secure a husband herself, her brother 
(the generous * syala ') would offer a dowry for ber marriage. 11 
The brother was indeed largely responsible 12 for the sister's 
.settlement in life; but besides providing a dowry in special 

1 Cf. Delbruck: lad- Ver., 574, 576, 582; Zimmer : Alt. Leb. 309 

(opp. but not clear) j but cf. Jaim. Upan. Bra. Ill, 12, 2. 

2 With this feature may be compared the comparatively greater pre- 

valence of widow-burning and conjugal fidelity amongst the ru ing 
nobility of the earliest times as shown by instances in ' tradition.* 

3 Bv. X. 85~ 46; Av. XIV, 2, 26; cf. Ait. Bra, IH. 37. 

4 Of. V.I., I, 527, 

5 Brhadd. V, 49 ff . 

6 Bv. X, 78, 4; 85. 15-23: vide n. 5 above. 

7 Bv. T, 123, 11 ; Av. II, 36 ; etc. 

8 Bv. V. 61, etc., with Brhadd. V, 49 ff. 

9 Bv. X, 34, 3. 

10 Bv. T,' 109. 2; VTTI, 2. 20; Matt, Sam- T. 10. 11; Tait'b. Sam- 

TI. 3, 4, 1: Kath. Sam. XXXVT. 5: Taitt. ' Bra. I, 1, 2, 4; 
(cf. Nir. TI. 9; Mann. "ill, 53; VIII. 204; IX, 98). 

11 Bv. VI. 28. 5; X, 27, 12; Av. V. 17. 12; Bv. I, 109, 2 f syala) ; 

X, 85 T 6 (anudeyi). 

12 Of. Zimmer: Alt, Leb, 328. 

( 98 ), 

cases he seems only to have exercised & general supervision 
over ' his sister's love-makings ,-f or it is considered a bad 
thine to take advantage of defenceless brotherless girls 1 : which 
shows that girls with brothers were to some extent guided in 
their social intercourse. Sometimes, again, & father could 
make a gift of his daughter to someone for services rendered, 
as in the case of Oyavana or Syavasva 2 ; or in special cases 
he could stipulate for his daughter's remaining with him after 
marriage and bearing sons for his family only. 3 ^ These facts 
show some amount of control over the daughter's marriage, 
who could, under exceptional circumstances, ' be sold/ given 
away in arranged marriage-, or bestowed as a gift ; but if the 
daughter liked, she could go definitely against her father's 
wishes, and be appreciated for that, as in the case of Eamadyu, 
daughter of Purumitra, who practically eloped with. Vimada. 4 

The so-called marriage Hymns are rather tesselated pieces 
(as already noted). 5 A number of features mentioned in 
them contradict one another, or do not fit in; though later 
Sutras have tried their best to use these passages to suit con- 
temporary ritual, often obviously misapplying them. 6 Thus 
some of them refer to the bride's being first escorted as a 
* kumari,' from her father's house to her future home in 
procession, where the marriage and its consummation takes 
place, 7 while others might refer to the * wedded s bride being 
so carried in procession 8 ; some refer to eager and favourable 
brides, 9 others, to wailings of the bride and other women in the 
house, 10 which evidently refers to a mock-ceremonial attend- 
ing the ' Baksasa * form of marriage by capture of a wailing 
woman 11 (supposed to make the bride and her sisters, friends 

1 Rv. ir, 5 3 5. 

2 Jaim. Bra. Ill, 12, 2; Brhadd. V, 49 ff. 

3 Rv. Ill, 31, 1. 

* By. I, 112, 19 ; 116, 1 ; 117, 20 ; X, 59, 7 ; 65, 12. 

5 Vide ante ; sec. re polyandry. 

ft As in generally applying evidently polyandric forms, with the absurd 
result that a normal wife is called ( devr-kama/ and ascribed 
several husbands and consummations ; or, as with the funeral 
mantras (vide sees, re widow-remarriage and ' smasanas'), -where 
passages relating to widow-burning are used of a cow, and those 
referring to erection of mounds are applied to digging out. 

7 Av, XIV, 1, 62-63, with XIV, 2, first part, sp. 6-19 ; probably also 

XXV. 1, 6-22. 

8 Av. XIV, 1, 61 ; 2, 74. 

9 Av. XIV, 1, 9.31 ; 2, 52 ; etc. 
10 Av. XIV, 2, 59-61. ' 

H The passage becomes perfectly intelligible and appropriate if * saina- 
nartisu' is analysed as e samana-rtisu * (* in wedding assembly 
combats * ) instead of * samd-nartisii * ( * co-dancings ' } with 
Whitney. It cannot very well be supposed that funeral mantras 
have been inserted in the midst of marriage fanmrtae. Neither 
WMtney and Roth's Index to the Av.> nor Bath's Die*, notices 
this important word. c Bti * =conib^t is a very common wor4 

and relatives, miserable) ; some can only ^ apply to polyandry, 
to polygamy, or to sister-marriage, 1 while others apply to 
normal forms ; and at least two of the passages^ show traces 
of having once been part of some older Ikavidian ritual. 2 
Hence it is not safe to take them as describing in detail and 
in order any -one form of standard marriage-ritual; though 
some of their features may: well have been common to all 
forms and constituted the "special act of marriage : like the 
taking of the bride's hand, 3 the circumambulation of the house- 
hold fire, 4 or the consummation before or after home-coming 
(with connected rites) . 5 

Apart from these optional forms of ritual (associated with 
extraordinary types of marriage and traces of different tribal 
customs) , more of variety have been introduced by the different 
manners in which marriages were settled. The part of the 

* bride-wooers ' in several passages of the marriage-hymns 
shows that often alliances were negotiated 6 by intermediaries 
(who were either friends and near relatives of the bridegroom, 
or professional match-makers) 7 ; yet, generally the bride her* 
self is approached and won over by favourable representations 
about her suitor, and .she eagerly approves of the match. 8 
Indeed in !Etgvedic opinion, 9 that * vadliu s alone is * bhadra,* 
who, brilliantly attired, herself selects her mate (' friend') 
even in the midst of an assembly, though it is at the same 

in Vedic texts. * Samanarti * thus accurately describes the con- 
flicts at Svayamvaras and forcible carrying oS of brides from the 
marriage -assembly, so amply illustrated in 'ancient * traditional ' 
accounts; from real wailings (with dishevelled hair) of the 
bride's ' janah ', ' jamis ' and other * yuvatis ', a formal _ ceremony 
would develop as a survival, and it is apparently this that is 
referred to in the above passage. 

1 Vide ante, sees, re polyandry, polygamy and sisto-marriage. 

2 Av. XIV, 2, 19 ; 63 ; in the former the foridfc as the new mistress of 

the house addresses the house or its spirit as * I$e % which can 
only stand for the Dravidian * i$a ' and cognate words meaning 
homestead; in the latter the bride scatters * pulya', which again 
represents the Dravidian pull ' (cf. PraErta ( pulla * and mod, 
vern. " mudi.* 

3 Av. XIV, 1, 51 ; Ev. X, 18, 8. 

4 In the Grh. Sutras. 

5 Which together take up a- large part of the Ev. and Av. marriage 


6 Av. XIV. 1, 8.9; 31 ; 2. 66; VI, 60, 1; etc. 

7 Thus in gyavasva's case the ' yara * was his father ; ^ while the 

* aryaman * who is busy finding ou wife for the wifeless and 
husband for the spinster (Av. VI, 60, 1) is evidently profes- 
sional 'ghataka.' In the Vats. Ka. SSt., the * varas ' are still 
near relatives or friends of the suitor, bnt the * varana ^ system 
is disparaged, preference being given to the s Gandharva, * where 

* varas * are needless. (Probably it is through the ( varana 
being thus often personally done by tfoe suitor that in later us 

* vara* has come to mean the bridegroom himself). 

* Av. XIV, 1. 8-9; 31; VX, 60, 2-3; etc. So also, the ' kanya adorns 

herself wilfc ornaments, eager to come to her ' vahattt ' i B*. -TV , 
58, 9. 

* BY. X, 27, 12. . . 

( 100 ) 

time recognized that many young women have to appear pleas- 
Ing unto suitors anxious to wed them for their fortunes. The 
most usual type of marriage-alliance seems to have been that 
in which the bride and bridegroom had previously come to 
enjoy one another's company, in their ordinary village life, 1 
or in various opportune festive gatherings/ and in which 
their free choice (made amongst a number of ^suitors and 
husband-seekers) and mutual attachment (growing through 
stages of estrangement, jealousies, wanderings and longings, 
and fostered by magic rites) had been approved as ; a, matter 
of course by their kinsmen, 3 who joined in the festivities 4 : a 
smooth and happy sort of affair with nothing rigid and un- 
natural in it. 5 But sometimes the lovers came into conflict 
with their guardians, and the marriage had to be accomplish- 
ed by capture and elopement, which was regarded as a com- 
mendable step for the knight and the lady alike ; thus in the 
case of Vimada and Purumitra's daughter, it appears that 
there was no violence pure and simple, 6 but that the affair 
was pre-arranged with the consent of the fair lady who refused " 
to be guided by her father. This previous mutual consent is 
a noteworthy fact, as being present both in marriages by 
' capture *" and those by 'gift.' 7 The gift of a maiden in 
marriage for services rendered is another exceptional form ; but 
other elements sometimes clothe its bareness : thus in the case 
of Hathaviti-Dalbhya's daughter, Syava^va was at the same 
time an ardent suitor for the maiden .subsequently * given ' to 
him. Priests very often received, from their princely patrons, 
noble maidens or slave-girls, for services at sacrifices, who are 
termed ' vadhu's (either wedded or ' wedable ' girls, or 
simply those 4 borne away' as presents on cars) 8 ; but this 
does not appear to have involved any proper' marriage, and is 
to be regarded as concubinage associated with polygamy, 
developing amongst certain opulent and powerful classes. 
Sometimes again, bargains were struck, and the bride was 

1 Vide pp. 95 and 96 above. 

a Vide pp. 94 and 95 above. 

3 E.g.^Av. II, 36 shows that parents usually left the daughter free 

in these respects,, and directly encouraged her in being forward 

in^ love affairs. (So also, even in her childhood her mother 

* thinks of the time when the daughter's developed youth 

f uativedinau ') wnuld win a husband for her). 

^ Ev. IV, 58, 9; Av. XIV, 2, 59. 

5 This ^is the type of marriage- alliance which, centuries later, is ape- 

dally recommended as the best form, and treated aa normally 
prevalent (in spite of the dicta of the law-books), in the Vats. 
Ka. Slit., under the technical name of ' gandharva-' This treatise 
closely follows the Velic notions about sex-relations, and re- 
presents conditions somewhat different from those in the kw- 

6 Marriages by forcible capture were of course Ipaown; vide ante. 

7 Such an element is also emphasised in the Vats. Ka. Sut. in tbe*e 

two forms. 
s Vide ante, sec. re polygamy. 

ioi y 

practically sold for a heavy price, or the bridegroom purchased 
by offer of dowry; but the former was considered discredit- 
able to the bridegroom, the latter creditable for the bride's 
relatives ; and both practices were resorted to in exceptional 
cases only, where, of the suitor and the bride, one had some 
undesirable defects. 

In agreement with the generally' free character of the Vedic 
marriage, is the absence of any great restrictions on marriage 
outside or within certain spheres. There is no ban on marriages 
within the same group of agnates and cognates; and the 
several classes, Aryan as well as Dasa, can intermarry. Sister- 
marriage, however, was apparently falling into disuse towards 
the close of the Itgvedic period 1 ; but even in the subsequent 
Brahmana period the restrictions on ' sagotra * and ' sapinda ' 
marriages did not go beyond the third or fourth generation on 
either side 2 ; and fir,st cousins, through mother 1 ' s brother or 
father's sister, could marry, 3 amongst several .sections of the 
people, marriage with a paternal uncle's daughter being more 
in use 4 ; the restrictions grow more and more marked later on 
in the Sutra period 5 ; it is thus quite clear that they amounted 
to very little in the Vedic age proper. 

So also, intermarriage between the several * vanjas ? was 
much easier. It is indeed inconceivable how young men and 
women could have been allowed free social intercourse in public 
gatherings or in private company, if there were any real bars 
to such intermarriage. This may have taken the form of 
hypergamy oftener. A Yv. Samhita, however, mentions the 
' ayogu,' 6 which^ if it is connected with the later ' ayogava/ 
may mean the Arya woman (vaisya) married to a Sudra 6 ; the 
evidently old tradition recorded in the Asvalayana Sutra, that 
equally with the ' devr,' the family slave (' dasa ') could law- 
fully marry the widow of his master, 7 is a clearer fact for the 
early Yedic period; the Yv. Saxnhitas also refer to frequent 
cases of Sudra-Arya connexions, 8 which points to the beginning 

1 Vide ante, sec. re sister-marriage. 

2 E.g. Sat. Bra. I, 8, 36. 

3 Of. the many Buranic as well as Buddhistic ca&es. 

4 Vide V.I., 1, 236. 

5 E.g. Gobh. Grh. Sut. Ill, 4, 4-5; Apast. Dh. Sut, II, 5, 11, 15, 

16; etc.; (for c sapincla * marriages; Gaut, DJi. Slit. XIV, 13; 
Vas. Dh. Sut. IV, 17-19; cf. Mann : III, 5; Yfcj, Dh, Sis, 
I, 52 3 53). la Vats. Sut. the maternal uncle's daughter is 
still frequently courted and married, amongst the Dki5.atyas= 
as well as elsewhere, wher the young man is more or lesa 
dependent on his maternal uncle or lives with him, 

6 Vaja. Sam. XXX , 5; Taitt. Bra. in, 4, 1, 1. In Furanic tradition 

Marutta, a famous Aiksvaka prince, is called an ' ayogava ' j 
'this family is said to have been degraded to the Yaisya status 
in ancient times owing to a mesalliance. (Query ; Can it- then 
be inferred that the Iksvakus were originally Sudras?), 

? Asval, gr. Siifc. IV, 2, 18. 

3 Vaja, Sam, XXIII, 30*31; Taitt. Sam. VII, 4, 19, W- 

( 102 ) 

of such intermixture in the earlier period ; in an Atharvavedic 
rliariQ directed against a rival lover (or a wife's paramour) he 
is referred to as a ' dasa,' winning her love by sheer physical 
strength. On the other hand, men of the Bgvedic priestly 
class are often stated to have married into royal families, as 
Cyavana, Syavasva, or Virnada did. 1 Probably this apparent 
prominence of -hypergamy is due to the notices coming from 
the brahmaijLS, who have naturally passed over l rajanyas ' who 
married brahmaia women : still there is the clear case of King 
Svanaya-Bhavayavya's beloved wife who was an Angirasl. a 
The Atharva-veda glorifies thd brahmaji as the best husband for 
women of all other * varnas/ 3 though from the same context 
it transpires that the ' brahman! ' often held opposite views, 
and had to be reclaimed from persons of other ' vamas ' 
with the help of the king's justice. 3 ' Vaisi-putra's" are 
known ^to the early Brahmanas 4 ; in the Yv. Arya-Sudra 
connexions are subjects of jest amongst court and priestly 
circles, 5 so that legal marriage between such must have been 
frequent; and respectable Vedic personages, like Ausiia 
Kavaea or Vatsa, were sons of slave (' dasl ' or ' sMra ') 
mothers. 6 The use of the term ' dasl/ as compared with that 
oi dasa/ in Vedic texts, shows that the ' dasl ' very early 
came into contact with Aryan masters, 7 as a result probably of 
toe extermination and subjugation of aboriginal tribesmen : 
accordingly dasi-putras * became quite common, and slave- 
girls presented to pnests by conquerors could be called 'vadhu's 
or wedable women. 8 

* am ifche Saryataa and the 
- . instances. Kakivant' s cas* 

mixed one on either side. 

'^Ti '^ c caaS ^"S- th afc ^ Yayati). 

( the sfcnkm g anecdote of Oghavati in the Epic. 

mStattCSS f aberration8 of%rihm a9 l s in IL 

, 7, 3; ^ Bra. 30U, 2. 

Vide note 8, page 101. 


> - Weber 
: Sans. Bead. 386-'87; 


\ide ante, sec. re siave-conWjinage 

Vide ante, ibid, 



It is significant that almost ail the Vedic terras denoting 
woman express a special sense of actual or potential wifehood, 
and very few carry the simple meaning pf woman &s opposed 
to man. In Ev. the latter sense is found undoubted in fc stri ?I 
(as opposed to ' purnarns ' or * vrsan ' and as watna-n general- 
ly) ; but from Av. onwards * stri ' comes to be opposed to 

* pati J and to mean wife, 2 though as late as the Sutras 4 stri ' 
is still distinct from * jaya,' and the general sense* of * woman ' 
always remained associated with it. * Mena ' 3 denotes, first, 
any female (of animals, etc.) , then a woman (but the sens of 
' potential motherhood ' may be implied).. ' Kana ' 4 and 
other cognate terms (' kanya/ 5 * kamnaka/ 6 * kanyana/ 7 
' kanyala * 8 ), in Ev. and Av., denote a 'maiden end young 
woman/ with no direct reference to wifehood, but are often 
used in contests showing her fitness for wooing and marriage. 

* Yuvati/ 9 while meaning ' youthful woman and a maiden/ 
implies a readiness for union with a * marya ?1 C young lover '), 
In some later Brahmanas 11 * yosa * hag sometimes the sense of a 

* girl 5 (in Av. as well), 12 or of ' female >13 generally, as opposed 
to * vrsan '*; but, though occasionally the Ev. has it in the sense 
of a * daughteiV 14 in the great majority of its occurrences 

* yosa ' ls means ' young women, specially maidens, as meet for 
wedlock/ while the Av. has also the sense of 'wife ' 16 ; the 
cognate terms ' yosan/ 17 'yosana 18 and f yosit/ 19 also, 

. I, 164, 16; V, 61, 8; etc. (also ia : Malt. Sam. IV, 7, 4j Taitt, 
Sam. VI, 5, 8, 2). 

2 Av. XII, 2, 39; cf. Ait. Bra. Ill, 22, L 

3 Ev., I, 62, 7 5 95, 6; II, 39, 2. 

4 Bv., X, 61, 5; etc. 

5 Rv. I, 123, 10; 161, 5; III, 23, 10; etc.; Av. I, 14, 2; XI, 5, 18; 

Xn, 1, 25, etc. 

6 Rv. IV, 32, 23; X, 40, 9; (Nir. IV, 15). 

7 Ev. VIII, 35, 5. 

, 54, 14; IV, 18, 8i V, 2, 1, 2; IX, 

86, 16: X, 30, 5; Av. XIV, 2 3 61. (This sens continues in the 
Bra. Taitt. Bra. Ill, 1, 1, 9; 2, 4; Sat. Bra. XIII, 1, 9, 6; 

4-, 3, 8; etc.), 
10 Rv. Ill, 31, 7; 33, 10} IV, 20, 5; IX, 96, 90; *tc. 

U It III: I'. I^XIV, 1, 56 ; etc.; VI, 101, 1, etc. s (cf. Ddbrfick : 

Ind. Ver., 413). 

13 afc. Bra- I, 2., 5, 15; (freq. m Bras). 
W Bv. I, 117, 20 j (cl Zimmer: Ait. Leb., 310)* 

15 Kv. I, 48, 5; 92, 11; HI, 33, 10; 38, 8; etc. 

16 Av. XII, 3, 29. 

1? Ev. TV, 5,*5. f , 

18 Rv. Ill, 52, 3; 56 } 5; 62. G; VIl f 95 5 3; etc. . ^ 

19 Bv. IS, 28, 4. 

( 104 )" 

have- in Bv. the import of ' woman &s young and ripe for 
marriage.' ' Narl 51 in Ev. has the 'dear import of woman as 
wife, as affected by matrimonial relations pati-justa/ 2 
; avidtava/ ' supatnih/ 5 etc.), though in some cases the 
sense of ' woman as the sexual complement of man ' is possible. 4 
The term ' gna ' probably meant { woman ' originally, but 
was Dearly restricted to 4 divine women ' ; but there too, these 
are 'wives * of gods. It is doubtful if ' jam ' (' jani ') has 
the general sense of * woman ' 5 or a derivative sense of 
' hetairai ' 6 at all in the Kv. or later : it is almost always ap- 
plied 7 in relation to * pati,' and phrases are used joining the 
word with husband, marriage, or wifely position (' patyur 
janitvam/ 8 ' janayo na patnih/ 9 etc.). 

It would thus appear that from early Viedic times the 
woman has mainly and almost exclusively been conceived of as 
wife and mother, and marriage was her normal and inevitable 
condition. This aspect of the woman, is emphasised in the 
terms applied to her as wife : she is ' jam ' a*s bearing her 
husband's child, 10 and ' jaya >u in the same sense along with 
that of the object of marital affection. Even as daughter she 
is 'duhitr/ 'the potential nourisher of a child.' 12 This 
characteristic conception of woman has determined 1 largely her 
place in the social system of Ancient India. Her special 
sphere is therefore the home; and she has always been true 
to it, though from the Vedic age onwards, at different times, 
she has passed out of her groove and lived a much fuller life. 15 

Such being the standpoint from which the woman, was 
regarded, it would seem to follow that a daughter was a 
welcome addition to the family, Though in Ev. the birth of 

1 Bv. VII, 20, 5; 55, 8; VIII, 77, 8; X, 18, 7; 86, 10-11; (also in 

AT. XIV, 2, 13; Vaja. Sam, XXIII, 36; Ait. Bra. Ill, 54). 

2 Bv. I, 73, 3. 

3 Bv. VII, 20, 5; X, 18, 7; etc-; (same sense later also, sometimes, 

e.g. (Jaut, Dh. Sut. IX, 28), 
Cf . Delbrack : Lad. Ver., 417.439. 

5 Bv. IV, 52, 1 (Usas, a fair ' jani '= wife?); V, 61, 3 (here 'wives ' 

is reqd., but cf. Delbruck ; op. cit., 413). 

6 Bv. J, 85, lj IV, 5, 5; 19, 5; VII, 18, 2; 26, 3; IX, 86, 32j (cf. 

X, 43, 1) ; (also in Vaja. Sam. XII, 35 ; XX, 40 ; 43 ; etc.) ; 
the plural use is no good ground for this sense. 

7 BT. X, 110, 5 ('patibhyo na janayalj ') ; VIII, 2, 42 ( 4 janit-vana ') ; 

V, 61, 3 and X, 40, 10 (r*l to married condition). 

8 Bv. X, 18, 8. 

9 Bv. I, 62, 10; 186, 7. 

10 Of. 'jam* contrasted with ' patni ' : Rv. I, 62, 10; 186, 7. 
H Frequent in Bv. and' Av. (tor refs. vide V.I. ? I, 285-6); cf. the 
distinction in use of tne terms * jaya f and ' patmV 

12 Cl Belbruck ; lad. Ver., 454; V.I., I, 371. 

13 E.g, in Bgvedic, TJpani$adic and Buddhistic periods. 

'( 105 ) 

sons is specially desired, 1 nothing is said in it deprecating that 

of daughters ; an ancient ' gatha ' cited in the Ait. Bra. 2 

apparently calls a son ' heavenly light ' and a daughter 

* misery/ but ' krpai?.am ' there might as well mean ' evoking 

tender feelings and compassion/ and a contrast is not required 

by the context ; but in tne Av. 3 female births are often regard- 

ed as unpopular, 4 being apparently the view of the common 

people (witii whose practices the* Av. was largely connected) : 

thus we hear of charms 5 for changing the foetus into a male 

one (the source of the later ' pmnsavana/j and of herbs which 

scared away demons seeking to convert it into a female. 

Female infanticide was, however, probably non-existent. 

Apparent references 6 to exposure of girl-infants may mean 

nothing more than ' laying aside the girl and taking up the 

boy/ 7 or ' getting rid of the girl by marriage 18 (though even 

this would imply that girls were not cherished). The very 

fact that later riarnhitas 9 (as well as Bra,s, Araiis and 

Upan.s) 10 severely condemn ' bhruna-hatya ' as the greatest 

crime would go against a. supposition 11 that female infanticide 

was a Yedic practice, though this condemnation refers to the 

' bhruna J only, whose sex is yet unknown, 12 and may not have 

applied to the k born * female infant ; it is to be noted in this 

connexion that exposure of infants on other grounds was not 

unknown : the child of an illegitimate union is abandoned and 

exposed, 13 and there is an old Brahmana reference to two infants 

(probably boys) being exposed by a t'ather 14 ; so that if female 

infanticide existed it would certainl have found clear mention. 

1 Ev. I, 91, 20; 92, 13; III. 1, 25 ; X, 85, 25, 41. 42, 45; Av. Ill 

23, 2 5 V, 25 3 11; VI, 11. 2; etc. 

2 Ait Bra. VII> 15. 

3 Av. VIII, 6, 25; and VI, 11, 3. 

4 On the other hand cf . Av. X, 8 ? 27 ; * -fchou art woman and man., 

boy, also girl ' (referring to human life as a mystic and divine 

5 Av. VI, 11; and III, 23. 

6 Kath. Sam XXVII, 9 ; cf . Taitt. Sam. VI, 5, 10, 5; Mait. Sain. 

"IV, 6," 4; 7, 9; Saokh. Sr. Sut. XV, 17, 12; (Nir. Ill, 4). 

7 Bohtlingk: Z.D.M.G., 44, 494-'96, 

8 Traditional rendering by comm. 

9 Taitt. Sam. VI, 5, 10, 2 and 3; Kath. Sam. XXVII, 9; XXXI, 7; 

Kaplsthaia Sam., XLI, 7; Mait. Sam. IV, 1, 9; cf. Av. VI, 112, 

3; 113, 2. 
10 Taitt. Bra., Ill, 2, 8, 11 and 12; Taitt. Aran,, II, 7, 3 and 83 

Brhad. Upan., IV, 1, 22; Kaus. Upan., Ill, 1; (Nir., VI, 27], 
n E.g. in Zimmer: Alt, Lob., 319-20; Delbriick ; lad. Ver., 575; 

Webed : Ind. Stud., 5, 54, 260, etc. 

12 Cf. Taitt. Sam., VI., 5, 10, 2-3 (guilt attaching to slaying an 

undiscriminated embryo), 

13 Rv. X, 99, 12 ; cf . Rv. IV, 19, 9 ; 30, 16.19 ; also I, 112, 8 ; n, 13, 

12; 15, 7; X, 61, 8. 

14 Pane. Bra. XI, 8, 8; Ynkta^va Angirasa, did it: hence sacred 

knowledge which departed from him had to bo regained by rites. 

( 106 ) 

When In spite of all prayers and spells it was after all a- 
girl who descended on the family, it appears that ,she was not 
ill-treated in any way ; for ' when a father and mother begat 
both son and daughter, the one engaged himself in the business 
of his father, whUe the other received honour n (and ' the son. 
less father ensuring his daughter's progeny lived content...... 

honoured Ms son-in-law and went to the son of Ms 

daughter s ) . The husband and wife, sacrificing together deem 
it- * a favour of the gods, if they reach t'heir full extent of life 
with SODB and daughters by their side.' 2 In a battle-song s 
while the bowstring whispers like a loving wife, the quiver is 
praised as the ' father of many daughters ' s (the point of the 
simile being, ' who as well as shafts overcome the hearts of 
men ') ; so, to be a- father of many daughters was not at all 
regarded as unlucky and its advantages were appreciated. 

A happy love-match for their girl is the greatest concern 
of her parents, and they try all sorts of natural and supernatural 
means for that end. 4 When the married daughter left her 
father's home, 5 the benediction pronounced w,as full of tender- 
ness (referring to the plucking of the fruit from its stalk and 
the untieing of Varuna's knot, the bond of parental affection). 
The parents of daughters were not very anxious to ' get rid of 
them by marriage' 6 ; though from the Av. it appears that 
charms were uttered to secure husbands 7 for their daughters, 
yet it was only to strengthen her own endeavours ; the match* 
maMng * bride-wooer ' was entertained, but his business was 
to win the fear of the, maiden herself. 8 The m'otheir would 
sometimes refuse to give her daughter to one not up to her 
ideal, 9 even when the father had no objection ; and she resents 10 
when her daughter suffers in the hands of a son-in-law addicted 
to gambling. 

The mother no doubt wanted the daughter to help her in 
household work, and the unmarried sisters in the family to- 
gether brought home water from the wells, 11 in jars poised on 
their heads ( c seen by everybody but not known by the mind') , 12 

1 Hv. Ill, a, 1-2. 

2 Ev. YIII, 31, 5-9 ( ( kuinarina ' and ( putrina * in v. 8), 
s Rv. VI, 75, 1-7. 

* E.g. adorning, sending to Samanas, instructing in arts, encouraging 
i *..JS7 e ' entertaining * bride-wooers/ performing magic rites. 

5 Av, XIV,, 1, 17-20 and 1 46, 

6 Vide XL 8, p. 105. 

7 Av. H, 36. 

8 Av. VI, 60, 1. 

9 Cf J?^ * s Xr? va e P isode * (Bv. V, 61, sp. ; and Rv. V, 52-61 ; 

in T, ^ I 111 ' 35 ~ 58 IX > 52 ^ B * hadd - V, 492.). 

iu w. Xj o4 } 3. 

11 Rv. I, 191, H (seven 'agrus'.. kumbKnl * ). 

w Av. X. 8. 14 fa beautiful simile, repeated often in later poetry). 

( 107 y 

and wove and embroidered garments, 1 for their own future 
husbands as well 2 ; but at the same time they were not crushed 
with! domestic duties, and could join the merriments of the 
village youths, with whom they swung in ' green and white * 
swings under the village banyan, with music of lute and 
cymbals and display o* peacock plumes 3 ; even as * yuvatis ' 
they bad leisure and liberty enough 1 to enjoy to the full the 
company of their lovers. 4 

The unmarried girl stays on with her father (mother or 
brother) for years together 6 without any resultant unpleasant- 
ness ; she is ironically described as sitting long with * the 
fathers/ 6 but that indicates the parental consideration she 
enjoyed 7 ; she, on her part again, looked to her father's interests, 
as Apala 8 cared for her father's fields (and his bald head). The 
very fact that home-staying old maids were not rare shows 
that daughters were not regarded by parents as undesirable 
burdens, though the daughters themselves would rather get 
married, 9 A 4 tanva ' 10 or c legitimate son of the 'body ' is said 
not to leave any share of the paternal property to his sister : 
this indicates that in the absence of such a ' tanva * the; daughter 
inherited or had preference over adopted or other sorts of sons. 
At any rate she was entitled to maintenance and maariage- 
dower 11 from even such a brother (who was also expected to find 
her a husband, 12 look after her social conduct, 13 and s along with 
his wife, 14 to guide h'er generally). In an age when adoption 
wa-s hated, 15 when daughters could stay on unmarried in their 
lather's house, till death without social penalty, 16 and when 
daughters' sons were thought as good as sons of the body, 17 the 

1 Cf. Rv, II, 3, 6 and Vaja, Sam. XX, 41 (two sisters embroidering 

j>esas on a stretched web) ; *Av, X, 7, 42 (two sisters weaving a 
web stretched on 6 * mayfikhas/ one drawing the threads, th& 
other setting them); cf. Taitt. Bra. II, 5, 5, 3 (same). 

2 Av. XIV, 2, 51. (the bridegroom wears this garment e soft to touch * 

in the marriage ritual). 

3 Av. IV, 37, 4. 

Rv. in, 31, 7; 33, 10; IV, 20, 5; IX, 96, 20; etc,; cf. II, 10, 5; 
(Nir. IH, 15; IV, 2). 

5 Rv. I, 117, 7; IT, 17, 7; X, 39, 3; 40, 5; Av, I, 14, 3; etc. 

6 Av. I. 14, 3. 

I An old maid was probably allowed to manage her father's household : 

hence the point of the remark ' Yama's kula^a ' ; cf . next note. 

8 Rv. VOT, 80. 

9 E.g. Ghoa, Apala ; cf. Rv, I, 117, 7 5 {cf. also the Av. charms, 

showing the girls' initiative in this matter). 
15 Rv. m, 31, 2. 

II Rv. I 3 109, 2. 

12 Cf. n. 11, and Zimmer : Alt, Leb. 328. 

13 Cf . Rv. I, 134, 7 ; W, 5, 5 : Av. I, 17, 1 (cf . Av. I 5 14, 2). 

14 Rv. X, 85, 46 ; Ait. Bra, III, 37, 5 (nndor wife). 

15 Ry. VII, & 7-S; (Nip, HI, 

16 Vide n. S above, 

17 Rv, IH, 31, 1. 


daughters' legal position acd importance in the family was 
evidently better than it was later on. Many daughters 
apparently inherited property in some way or other : for ' many 
a maid was pleasing to the suitor who fain would marry for 
her splendid riches/ 1 

As she grows up, the daughter is allowed a larger share 

of personal and social freedom 2 ; she is not rigidly secluded from 

the outside world, or hedged round with prohibitions. From 

sharing in the Tillage dances and swings 3 she passes on to 

constant companionship with her chosen lovers. 4 She goes to 

festivals, 5 adorning herself in desire ot marriage, 6 where &he 

may even spend the night, 7 She receives suitors 8 quite as an 

independent person; goes to trysts to meet her love, 9 or meets 

him in her own home 10 while her people are asleep ; she chooses 

her ( friend ' as her husband in the midst of assembled men 11 ; 

she may even elope with & knightly lover 12 against the wishos 

of her lather. Sne candidly tries to get a husband herself, 15 

before she becomes a confirmed old maid ' and it becomes too 

late/ 14 Brothers quite naturally exercised .some amount of 

control over the social activities of the young maiden, 15 but only 

to the extent of seeing that no evil-minded man took any undue 

advantage of them. 16 As brothers were normally expected 17 to 

be on the look-out for a match for the .sister, brotherless girls 

had often to be very forward, ' turning boldly towards men/ 18 

attracting attention by red garments. And in spite of some 

amount of social feeling against breaking the order of seniority 

in matrimony, 19 younger sisters were not wanting who were 

"anxious to woo' 20 before their elder sisters, and found husbands 

1 Ev. X. 27, 12 (*kiyaii yosa maryato vadhuyoh pariprlta panyasa 

varvena '). 

2 Cf. V.L, II, 485. 

3 Gf. n. 3, p. 107. 

4 Cf. n. 4, p. 107. 

5 Bv. IV, 58, 8; VI, 75, 4; VII, 2, 63 X, 86, 10; Av. II, 36, 1, 

6 Bv. Til, 2 5 5; I, 123, 11 - Av. II, 36, 1; cf. Bv. IV, 58, 9. 

7 For Samanas often lasted all night: Bv. I, 48, 6j VII, 9, 4- 

8 Rv. X, 27, 12 ; cf . n. 4, p. 107 and Bv. X, 30, 6. 

9 Ev. X, 34, 5; 40. 6. 

10 Rv. VII, 55, 5-8; cf. I, 134, 3; Av. W, 5. 

11 Bv. X, 27, 12. 

12 Bv. 1/112, 19; 116, 1; 117, 20; X, 39, 7; 65, 12. 

IS Rv. VXI 2, 5; Ar. n, 36, Iff.; cf. the ' didliisii 3 and ' agre-didEI$u.* 

14 Rv. I, 117, 7 5 ,39, 6. 

15 Cf. notes 13 and 14 p. 107: 
15 .Rv. IV, 5, 5. 

1? Cf. Zimmer: Alt. Leb. 328. 

18 Bv. I, 124, 4.7 j etc. 

19 Censured as - sinful In later Samlutas and Brahma^as> fallowed by 

Dhanna-Sutras (see n. 20 below and n.- 1, p. 109), 

20 Vas. Dh. Sut, XX, 7 ff. (cf. Vaja. Sam. XXX.9). Probably also 

referred to in Taitt. Sam. Ill, 2, 4, 4; Kat. Sr. gut. II 3 1, 22; 
l Sut, 3, 5j 137, 37. Tide note 1 : nest page. 

.( 109 ) 

brave enough to face denunciations or opprobrious epithets. 
Vedic society thus appears to have taken it for granted that 
the woman had her likes and dislikes, her loves and Joys, as 
much as the man. This personal freedom of action of the 
unmarried -woman develops into a dignified wifehood after her 
marriage. 2 

Both as wife and as daughter, women were admitted to 
the privileges of the highest education, at least amongst the 
intellectual sections of the people. 5 The early Vedic literature, 
as is well-known, contains contributions from women*; and 
women played an important part in the later Vedic period, in 
the Upanisadic discussions, 5 a fact which explains the subse- 
quent activities of women in the age of the Buddhistic Refor- 
mation. 6 In the society of the TV. Samhitas and Brahma^as 7 
women love music and marry by preference men who can sing, 
so that they must have ordinarily been taugEt dancing ^and 
music; thus ' gathas ' were sung at weddings, and in TV. ritual 
also the ' patnl-samans, or wives ' songs have a recognized 
position. IB an Upanisadic household it was thought worth 
while to go through, special ceremonials in order to secure _ the 
birth of a daughter who would distinguish herself 8 by learning. 
Learned women are often referred to in the Brahmanas, Upani- 
sa'ds and Sutras. 9 The Ath&rvaveda, in the verses in praise of 
Vedic studentship, declares that it is by virtue of her ' brahma- 
carya ' that a young maiden gets a husband 10 : this may point 
to some otherwise undetailed traditional course of instruction 
k> girls, 11 similar to the well-known system of schooling going 
by that name ; or it may well have 'been the case, that girl 

* E.g. < agre-dadhns ' (Tv. Sam's); agre-didhisu * (Yv. Sais, Taltt. 
Bra and. Dh. Silt's) - agre-didhisu-pati ' (Yr. and ^; ^ ; 
cf. k didhM-pati ' (Dh. Sut.) ref. to elder sister ; and pamrt ta 
and ' parivividana (in Av,, TV. Sams> and Bra. , rei. w 
breaking of order of seniority amongst brothers). 

2 When she is free, for instance, to address councils; vide ante, pp. B, 

10 and 11. , , , A iio-sn 

3 Cf. Hopkins. J. Am, Or, S., 13, 351- J 52: Weber : Ind. Stud. 10, 118-19. 

4 E.g. Rv V, 28; Vm, 80; X, 39j 40; etc. . TT 
6 E.g. Brhad. Upan. HI, 6, 1 8, 1 3 ital. Gf- ? ut I 

6 As evidenced in convents, missions, philanthropic and 

work. x, - T> - TTT 

7 Taitt. S'am^ VT, 1, 6, 5; Mait. Sam. Ill, 7, 3; etc. ; Sac. B :ra. HI, 

2, 4, 3-6 (where however music seems to be regarded as ratner 
a vain pursuit for man, s^iitin^ women better). 

f - Brhad. Upan. VT, 4, 17 (a ' ( pandita dnhit-a 1 ). TTT T 1 . 7 i , 

9 Aifc. Bra. V, 29; Kans. Bra. IL 9; Brhad. tJpan. III. ^ *> ^ L 
A^val. Gr. Siit. ITX. 4, 4; Safikh/ Gy. S3*. I \>^' , d 
Av. XT, 5, 18; (' hrafama-vadini ' women, amongst both io&ai an a 

10 , , , ^ 

priestly families, occur in Purlnic traditional accontita t 10 
very earliest steps; a few of them 'are mentioned in \eciic 

tuffe als^. e.g. Mamata-AngirasI). , . -, v& 

Courses of sacred instruction, for both boys and g iris are 
amongst many primitive or ancient tribes. 

< 110 ). 

students sometimes resided with the family of a- teacher for a 
number of years, equally with boy-students, a system implied 
in the Eplc-Puranle and in classical Sanskrit literature 1 as well. 
The extensire use of metronymics in post-Yedic literature 
{appearing from even the Egvedic times onwards) , 2 is partly 
accounted for by the fact that women of the more intellectual 
groups amongst the brahmans or ksatriyas had often as much 
reputation in the learned circles of teachers as their men, 3 and 
a metronymic must often have been something to be prond of, 
serving as a good introduction to its bearer (like * Gargi- 
puka*}. 4 Post-Vedic literature indeed knows of quite a number 
of women-teachers of philosophy and ritual, married or other- 
wise, 5 who apparently flourished towards the end of the Egvedic 
period and immediately after it. 6 " The unmarried (' kumarl ') 
women-teachers were designated gandharva-grhita/ or 
1 married to the Gandharva(s).' 7 

1 E.g., the case of Amba residing as a student with the Saikbavatyas > 

in the Epic; or the heroine of Kalidasa's famous drama, along 
with her friends, in the charge of the venerable matron of the 
hermitage. {The ref. herd may however b to purely Epic 

2 Vide ante. 

3 J5.g, Pataficala-Kapya's wife and daughter, Yajnavalkya's wives, etc. - 

Yajfiavalkya proves his superiority by showing that he knows 
all that the former two ladies knew; some of these women are 
included iu lists of rsis and teachers regularly honoured by 
Vedic students. Vide n. 5 below. 

4 Bthad. TJpan. 71, 4, 30. (01 the Vedic and post- Vedic metronymics 

some at least may thns refer to descent from women-teachers). 

5 Ait. Bra. V, 29; Kaus. Bra. II, 9; authoritative opinion of a 

'kmnari gandharva-grhltay on Agnihotra ritual. Pataiicala-Kapya's 
daughter _was_ a * gandharva-grhlta * : Brhad. Upan. Ill, 3 3 1 ; 
so was his wife: ibid HI, 7, 1; they instruct enquirers from 
distant lands; Fataficala himself learns from his wife. Ga-rgi 
Vacaknavi, Vadava-Pratitheyl and Sulabha-Maitreyi are classed 
with |-3is in tie Sutras ; cf. Sankh. Grh. Sut. IV. 10: Asm!. Grh. 
Siit. in, 4, 4. 

S Th first two references in n. 5 above relate to the time of a 
Jat-iiarnyaj the others refer to 'she times of Uddalaka-Zruni and 
* Yajnavalkya,' between two or four to seven generations after the 
Bgvedic compilation. It may be noted that PataHcala was an 
inhabitant of Madra, while the other names may be located 
in Mithila. 

7 Cf. V.I.^I, 486; with the exception of Pata^cala-Kapya's ' bharya ' 
who is also o called : apparently she was originally a ' gsnd- 
Imrra-grMta knmSri,* and had established her reputation as such 
before she married Pataficala, so that she continued to be known 
by 1 her old designation (or ' bharya * here may be tak^n in the 
older sense of ' female member of tfhe household,* ie., the same 
as PafcaHeala's ' daughter * mentioned in the same connexion). It 
seems (from *the context) that such women-teachers were supposed 
to be possessed by the spirits of ancient Angirala (or Atharv&me) 
seers. a remarkable point. 


This epithet is significant, and throws some light on the 

later 1 practice of formal or nominal marriage of courtesans or 

* artistes ' 2 to some deity or woodland spirit 3 ; it also explains 
the paradoxical statement in the Vedic marriage hymns, that 
three divinities are the first three husbands of a maiden, the 
fourth being the husband proper/ 4 Evidently the Yedic 
society -conceived of girl-life as developing through three stages 
(physical, moral and intellectual) into the fourth, 5 that of actual 
wifehood, where girlhood ended : the stage presided over by 
Soma represents gradual acquisition of beauty and grace, 6 that 
by Agni, of knowledge of domestic religious custom 7 and purity 
of character, and that by the Gandharva, 8 of various accomplish- 
ments. It follows that in theory every girl was supposed to 
have passed through a period of training and acquired some 
accomplishments, they may have been any thing from dancing 9 
to the subtlest ritualistic or esoteric doctrines 10 Sefore she could 

1 But. probably a very ancient practice ; marriage to a tree is known 

in the Jatakas. In the Av. women are "believed to be possessed 
and enjoyed by G-andharvas, apparently in tke course of village 
dances, music "and swingings ; probably the confirmed flirts and 
mnsical experts, who formed the central figures of village festi- 
vities, and refused to marry, were the first 'gandharva-grhltas, 1 ' 
They probably represent the ' apsarases ' of Yedic and Epic- 
Puranic tradition and the ' ganika's * of Buddhist and post- 
Mamryan periods; cf. their eminent position in the learned, 
literary and court circles as described in the Vats. Ka. Sut. 

2 Sometimes women of considerable wit and attainments, attached to 

the stage or the temple. Vide n. 1 above. 

3 The temple god, a Kumara image, or some tree, etc. 

4 Bv. X, 85, 40.41 =Av. XIV, 2, 3.4; cf. Av. V, 17, 2, 

5 The analogy of the ' asrania ' theory is significant ; probably it 

indicates an occasionally followed scheme of female education. ^ 
> Cf . the traditional comparison of a girl's development with the moon's 
waxing (e.g., in Kumara : I; cf. also the term 'sodas!/ which 
alludes to the 16 lunar phases). Soma might also signify, more 
particularly, the development ot adolescence (owing to the Mooa a 
supposed connexion with menstruation). 

The ref . in Av. II, 36. however, to ' King Soma making the maiden 
of good fortune * and to Soma and Brahman enjoying (tasting), 
and Aryaman enriching (renewing) her fortune (or youth, 
person), snggests another distinct yet similar conception 
(in perhaps another <age or society), according to which 
the King (typified by the legendary ancestor of all 
ilia ruling families), and the Brahman or H%h Priest of the 
tribe (or the priesthood as a body), were regarded as m theory 
(or perhaps optionally in practice) the legal ' masters * of every 
maiden of the tribe, till her marriage, which was supposed to 
be due to the good offices of Aryaman and favour of Agm : all 
this ace. to the divine law of Dhatar. The explanation of the 
comm. that Brahman =Gandharva (!) and so the ref. is to -2LLV, 
2, 3.4, is by no means convincing. . , . t_ 

7 Cf the vital importance of the wife for the fire-ritual in a house- 

' hold. Asni's lordship might also imply a period of 'brahma- 
carva' for the sake of suitable marriage 

8 The presiding gemns of the Fine Arts, like the Muses ; just- as the 

Apsarase^ patronized games and sport (Av. X, 1Q> 3), 
S Of. note 7, page 109. 
*0 Cf. note 9, p. 109; and note 4 5 p. HO, 

{ 112 } 

enter married life. 1 At the same time such entry did not put 
a stop to the activities of her preceding life-stages, as many of 
the women teachers and debaters were wives, 2 and could follow 
their husbands through all the stages of their intellectual and 
spiritual development. 3 It is also significant that in the Vedic 
society every woman .seems to have been conceived of as ever 
in a state of marriage, 4 as a child, with Soma or some other 
deity of abstractions, as a young maiden, with the Aits per- 
sonified, and then finally with her human husband , for whom 

- f 

indeed her mother impatiently watches the development 5 of 
her youth, carefully guides her toilet, 6 and for whom .she her- 
self weaves the soft nuptial robes in sweet anticipation, 7 

1 For in theory the husband is the ' fourth ' possessor of a- woman. 

2 E.g. Garglj Patancala-Kapya's wife; etc. 

5 E.g. Yajfiavalkya's wife; (the Vedic wife, like Mudgalani-Indrasena, 
could also stare the husband's martial glory). 

4 Cf. immediate remarriage or devr-marriage after widowhood. Of. 

also the later and modified doctrine ol Mann, regarding the 
perpetual dependence of woman on man. 

5 Av. VIII, 6, 1. 

6 Ev, I, 123, 11, etc. 

7 Av. XIV. 2. 51. 


Be Primitive Forms and Special Customs. 


There is a good deal of agreement between the evidence 
of the Vedic literature and that of the Puranic and Epic 
sources, with regard to the types of marriage, traces of its 
primitive forms, and the .general position of women in society. 
This is only what might be expected. In the scale of histori- 
cal values the Vedas and the rest of the priestly literature are 
still taken to be the standard, and whatever is not mentioned 
therein is taken to be non est or late and fabricated, while 
the least suspicion of a- mention is developed into an ingenious 
theory, often by tho same process whereby the sesasum of 
proverb changes into a palm-fruit. It is ignored that what- 
ever authority the priestly literature may have in questions 
of religious, mythological and theological developments (and 
even there it is by no means an exclusive authority) / it can- 
not, in the nature of things be taken as the prime and best 
source of historical facts. As is well known, priesthoods have, 
quite naturally, a strong tendency towards conceited isolation 
resulting in ignorance or ignoring of secular thought and 
events and towards perversion, of whatever knowledge of affairs 
they might acquire, to serve the interests of their own order 
and pretensions ; the first characteristic is displayed throughout 
the Vedic literature in both forms ; the second becomes notorious 
in the Puranic and Epic literature, the custody of which y 
according to well-atfcested traditions passed to the priesthood 2 
from the professional chroniclers and bardic experts, some 
little time after the catastrophe of the Bharata or all-India 
war, which apparently introduced a period of decline in the 
* Vedic ' ruling classes and court life, that had hitherto 
sustained this latter stream 2 of historico-literary productions. 
But even the mis-use of this sacred custody has not been able 
to obliterate the traditions of that early pre-Bharata age, 
some of which were too deeply rooted in the popular memory 

3 Gf . Sorensen : preaniable to the Index, for the growing conviction 
that Yedic religion and mythology cannot be properly under- 
stood without reference to Epic and Puranic. 

2 Cf. Chand. Upan. Ill, 4, where the King's daughter refers to 

herself as the daughter of the lauded person^ and the purohita's 
daughter as the daughter of the laudator, and so inferior. By 
this time therefore the Puranic chronicles had passed under 
priestly control from Sutas, and the time agrees perfectly with 
what the I*uranas themselves disclose. 

3 Distinct and independent, and associated with special classes arid 


( 114 ) 

and knowledge to be removed or wholly modified, even though 
offending against the priestly theories or subsequently changed 
ideas ; and through the blurring daubs and confusions of subse- 
quent brahmanical accretions and perversions, can still be 
discerned, thanks to the naive, uncritical, and unhistorical 
treatment of their otherwise intellectual authors, something 
of the original basic fabric. This supplies what is wanting 
in the Rgveda and other Samhitas and Brahmanas, namely, 
prima facie and bona fide historical events and conditions for 
most of the period covered by the former group. The value 
of this source becomes greater, when s incidental ' evidence 
in the * priestly * group of texts finds explanation, illustration, 
or support in the ' bardic ' one. 

The establishment of the position taken up here would 
involve a detailed examination of the historical elements in 
the entire Vedic, Puranic and Epic literature, a matter out- 
side the scope of the present dissertation. It will be sufficient 
to note here, that after a careful sifting of evidently later 
and brahmanical modifications, and rejection of all of those 
well-known extravagances of fancy , there still remains a 
residuum of fact, which cannot be given any other name besides 
4 traditional history, 3 which has every mark of having at 
one time been carefully handed down through professional 
recorders, and which can He given a tentative, workable, 
framework of chronology to stand upon, by a consideration 
and collation of undoubted synchronisms and uniform asser- 
tions. These synchronisms, plain statements, and the result- 
ant scheme of chronology, elucidate much ill-understood 
matter in the Ye die literature, correct wrong perspectives and 
give them their proper setting and importance. ' At the same 
time there is nothing in this clarified tradition that is really 
inconsistent with definitely 'Vedic ' facts. It is indeed strange 
that such an obvious source of historical information has so 
long lain outside the critical ken of scholars, and that so 
mpch of fanciful speculations, unnecessary theories, precon- 
ceived notions, almost prejudices, should have gathered round 
the study of that other group of texts, historically the most 
unpromising. But a wider comparative study -and estimate 
is bound to come, and a reaction is overdue. Often scholars 
shrink from it, as from an impossible task or perilous venture, 
simply because they have been accustomed only to the usual 
' Yedic ' studies conducted in a peculiarly bookish manner, 
and have imbibed the ' brah'manic tradition ' (if any) un- 
consciously or in spite of themselves. One has, however 
onlv to swerve the searchlight of critical study from ' Vedic ' 
to ' bardic ' lore, for a time, and then to a<jid fro, to strike 
the right course. As it is, we have too long been making for 
various misty uncertain shores, for the solar or 


myths, or the vegetation dramas 1 ; or been engaged, in 
exposing imaginary fabricators of tales from sacred texts, 2 in 
following the Indo-Afghan Vedic conquerors^ as they issued 
through the Khyber Pass, severed from their Persian kin, 3 
or in depicting the typical V-edic King, 4 strengthened in Indra's 
favour by the medicine-man, killing 99 noseless Dasas a day, 
ploughing his Punjab submontane field, tending his sheep and 
cattle, squatting on grass-mats, and sleeping in his hedge-girt 
hut or cow-pen, safe from forest spirits. 

The very fact that the traditional ' material makes 
clearer and fuller what might be obscurely suggested by the 
Vedic, 5 and some times vice versa, 6 and that a rational 

continuous history, dynastic as well as cultural, discloses itself 
on putting the two together, which sufficiently explains 7 all 
that is yet known about early Indian conditions, is a strong 
proof of the validity of the position set forth above. 

The results obtained from this view will now be detailed, 
so far as the selected topics are concerned. 

1 It will be enough to mention Ludwig's identification of Krsna and five 

Pandavas with the Earth and five seasons, and Keith's notion that 
the story of Krsna and Kamsa is a vegetation myth, which was 
often dramatised ritually. 

2 For this view cf. the recent Vedic Index. 

3 Even the recently discovered Boghaz-kui inscriptions have been sought 

to be explained away owing -to this preconceived notion. 

4 It is a common mistake *to take the Yedic period as a very short 

one and at the same time the most primitive one in Ancient India. 

5 For instance, the full explanation that the Epic-Pnranic traditions 

give, of the vague mentions of Kuril, PafLcala, and their kings, 
in the Bv., and Bra.s. 

6 As in the case of Dlrghatamas and Kaksivant. 

7 E.g., a rational explanation of Aryan expansion, of the _Inner and 

Outer Aryan groups, or o development of Brahmanism in the 
Sarasvfflfci and Kuru-Pancala country is a^orded by traditional 



As we have seen, sister-marriage was not very rare in the 
Bgvedic period (the references indicating its actual occurrence , 
and theoretic discouragement in the latter part of it) , The 
dynastic accounts in the common Purandc tradition, referring to 
the ruling nobility as well as the priesthood in that connection, 
contain many plain indications of the frequent occurrence of 
such consanguinous marriages, intermittently throughout the 
whole period covered by that tradition, viz., 90 steps, roundly, 1 
backwards from the Bharata War and the compilation of the 
Yedic texts. When these instances (along with those of 
other types and forms of sex-relations) are referred to and 
located in the general scheme of dynastic sequences, that 
evolves readily out of the patent synchronisms and consistent 
assertions, they 2 become very significant from the standpoint 
of early social history. 

The first instance of a sister-marriage in the dynastic 
lists is that of Anga and Ms ( father's daughter ' Sunltha, 
the parents 3 of the famous Vena. 4 As with other similar cases, 
the designation * pifr-kanya,' 5 though preserved without 

1 The Puranic tradition indeed goes back to still earlier times, and 

the^Ailas and Aiksvakas are treated as continuations of an 
earlier ruling race or races, portions of whose story are as much 
historical in form as the later dynastic accounts; some traces of 

the pre-Aila marriage-relations will be shown infra. 

2 The following instances are given in order of chronological sequence 

only, and not according to clarity of illustration. 

3 In all ^acconnts of Prthu-Vainya's ancestry in the Puranas and the 


4 Celebrated in Puranic texts (as well as in early Vedic texts); cf, 

"chosen King, an ideal one, supplanting ^prajapatis/ before 
.Nahusa, in the beginning of the (present) Yaivasvata epoch" : 
Padma : II, 35. 

6 This apparently curious expression becomes fully intelligible when 
it is considered that in the genealogical s'lokas it is the practice 
to describe a wife as so-ard-so's daughter, so that the only wav 
in_ which a sister-marriage could be described was to call the 
wife * pitr-kanya.' It ia possible that this expression was chosen 
as including half-sisters also, who would be only the father's 
daughters. In this connection it is noteworthy that in early 
Vedic texts (and the original Suta-Magadha texts' must have 
been_ equally ancient) * bhagini * does not occur, and ' svasr ' is 
a wide, general and relative term, while to designate sister as 
a blood-relation the qualification * jam; ' is used (vide ante). It 
is probable that * pitamaha-suta ' (or daughter of Prajapati or 
Brahma) in many genealogies really stands rfor a first cousin 
just, as 'pitr-fcinya *= jdster-. 

( 117 }, 

comments in one Purana, 1 lias given rise to emended readings 
and fanciful fables 2 : thus ' Mrtyu-kanya,' 3 is another read- 
ing for * pitr-kanya/ 4 which is closely connected with that 
figment about the mind-born daughters of the Pitrs.' 5 A 
Piiranic account also professes to give details of the wooing of 
' Mrtyu-kanya ' Sunltha, 6 where it is she who takes the initia- 
tive" in it; it is interesting to compare Yamfs similar attitude 7 
in the Yedic poem ; some of the later cases 8 also imply similar 
initiative on the part of the ' pitr-kanya/ viz. with Acchoda 
(m, Amavasu) and Narmada (m. Purukutsa). 9 

Eight generations after Mga and Sunitha's timte (accord- 
ing to the Puranic computation) we come across with several 
alleged .sister-marriages, amongst the Afresh groups of kindred 
races 10 that succeeded the Prthu-ites. 11 The clearest notice is 
that of Danu's son Vipracitti (by Kasyapa) marrying Diti's 
daughter Simhika (also by Kasyapa^, Danu and Biti being 

1 Matsya : 4, 43-44 (Svayambhuva Manu's dynasty, step No. 9). 

2 In most Puranas. evidently by way of after-thought or through mis- 


3 Matsya : 10, 3. 

4 Or * pituh/ or possibly ' Uroh ' or 'Muroli ' (the ' m r belonging to 

the preceding line) kanya. Urn being Anga's father. 

5 As Mrtyu=Yama=lord of the Pitrs. Cf. the brahmanical * pifcr- 

vamsa * sections of Pnranas; and Pargiter : AIHT. pp. 69-70; 
86; 196, 213. 

6 Padma : II, 29-35 : urged by her father. and_ helped in her plans 

by her companions* she arranged a meeting -with Anga (who 
wanted a strong successor), married him, and by him had the 
son Vena- . 

7 Her pl*ea of the necessity of begetting a worthy grandson for t-hezr 

father and her arranging to meet the brother suitably. 

S Vide infra* 

9 Probably this points to a type of sister-marriage similar to what. 

prevailed amongst- the ancient Egyptian ^ruling classes, where in 
the cnstomary consanguinous royal marriages the sister was the 
central figure. (Of. the dynastic history of Ancient^ Egypt, and 
the position of Cleopatra even in a much later period.) 

10 I.e., 'the descendants of Daks.a's daughters,' The Pnranic accounts 

of these pre-Aila races are well worth studying from the ethno- 
logical and geographical points of view ; they^ are consistent- in 
many respects, and seem to embody real racial memories. 

11 Some real personages of these groups have, however, become semi- 

mythical (e.g., the Panu-ite Vipracitti or the Vaivasvata Yama), 
apparently because- subseq-iDent developments of Aila and Aiksviks 
dynastic histories had little continued connections with these 
branches (after Bnsyanta in the Aila section, and earlier in the 
Aiksvaka section), and these, by dropping out of the chronicles, 
tended to become legendary.* But this does not make apy 
difference here; it is sufficient that such marriage-relations are 
indicated by tradition at this particular stage of traditional 

* Still^ven^n very much later times 3 the Dann-ite and Pf&f** Ponces 
of -traditional accounts are real persons diatinguahed from the 
mythical *s * maausya- dharmah ' or ' dhanyah (Vayu : 68, 
15 y i6VBrahmanda: III, 6, 1-3; etc.), probably they had some 
traces of non-Aila or non-Aik^vaka descent, though not always 

sisters and co-wives 1 ; it is to be noted that their descenflants 
(though recognized as a mixed " Daitya-Danava " clan) were 
called Sainihikeyas, after the sister-wife. 2 The Yama and 
Yam! of Bgvedic tradition are assigned by Puranas to the next 
generation, being children of Vivasvant, 3 one of Vipracitti 
and Simhika's step-brothers. Manu, another son of Vivasvant, 
also seems to have had a sister-wife : for Sraddha is stated to 
have been a daughter of Vivasvant, 4 and the genealogies make 
Sraddha Mann's wife; Manu, again, is called w Sraddha- 
deva ' 5 ; this ancient incest ascribed to a great name may 
have given rise to the Pnranic question : ' ' Why was Mann 
called Sraddhadeva ' ' which has introduced so many BraE- 
manical fables and didactic matter in the Pura^as. 6 But a 
more historical reference is to be found in the story of Cyavana- 
Bhargava, 7 (contemporary with Saryati-Manava, a step lower), 
who was the son of a Puloma, whom her previously k betrothed 
husband/ a Puloman, forcibly abducted from her * de jure ' 
husband Bhrgu's house : when the sacrificial Agni is said to 

1 Vayu: 67, 60; Brahmancla ; III, 5, 12: Hariv. : 3 3 184-'5; 204- J 5; 
213-'14; Matsya : 6, 25. Amongst Diti's near descendants, again, 

the Halahala ' gana ' (2 steps after Simhika) are said to have 
sprung from Anuhlada's son Vayu and daughter Sinibali : ap- 
parently another instance in the same group (Vayu : 67, 75 ; 
Brabmanda : III, 5 ? 33 tf.). 

Vayu: 63, 17-22; Matsya; 6, 25; Brahmanda : III, 6, 17-22. So 
also, otber branch races of this age are designated by metrony- 
micSj except the Vaivasvatas or Manvas, which may have an 
ethnic significance. But the point to be noted here is that the 
* mother-sid-e ' is stronger even in case of a brother-sister marriage. 

3 Son of Aditi, and alleged progenitor of the Aiksvaka (and Aila) 

dynasties. This bordering on myth need not be grained out, for 
real men. and women with names of favourite gods and goddesses 
have been very common in India; so in detailed genealogies like 
this, apparently reasonable traditions must be given their due. 
The reference (in the ' Aditya ' genealogies) to another contem- 
porary parallel of Vipracitti and Sinihika's case, in ' Indra ' 
son of Adifci and his wife lSaci-Paulomi 3 may be legendary; never- 
theless 'the traditional ascription of consanguinous connections 
to several members of a group has some value. It is curious 
that Piisan, who is a brother of Indra in these Puranic tables, 
should also be described in the Ev- as wooing his sister (vide 

4 Mbh. XII, 265, 9449. 

5 Mbh. XII. 4507; bat in XII, 13219, graddha-deva= Vivasvant (pro- 

bably wrong for Vaivasvata?). 

fi Of. Hariv. 16-18. It is to be noted that the Puranic tradition assigns 
the origin of the cnlt and ritual of ' Sraddha ' from compara- 
tively later periods, either from the time of Nimi son of Datta- 
treyaj or from that of Jamadagni, both ascriptions relating 
practically to the same age, much later than Manu*s. So the 
orahmanical connection between Manu and ' graddha ' is wrong 
and probably dates from after the standardization of Manu's 
code, by which time an explanation of Manu's incest had become 
necessary ; * Sraddha-deya ' is therefore derived from his wife 
and sister Sraddha, just as Rama has a Variant appellation 
Sitapati : (probably * &raddha-deva * would be a better reading). 
I Mbh. 3) (Puloma) : \ 5-7. 

( 119 ) 

have admitted his rights over her (she being his by choice, and 
Bhrgu's by formal rites). This seems to refer to a custom 
among the Pulomites (cognate to the DIti-Ites) 1 of 
consanguinous marriage, probably a brother-sister one. 

Two steps further down we come upon firmer ground, and 
henceforwards the references are without doubt historical in 
character, the details being dynastic and incidental. 2 The 
famous Nahusa-Aila is stated to have married a ' pitr-kanya,* 
Viraja, 3 who became the mother of Yayati, etc. In the same 
connection Amaysusu-Aila is also stated to have been chosen bv 

k> r 

4 pitr-kanya * Acchoda 4 as her husband, apparently in the face 
of some opposition. 5 So Nahusa had before him the precedent 
of his paternal uncle (the founder of the Kanyakubja line). 
In the same generation as Nahusa' s, and in the same part of 
the country, 6 there was another clear case, amongst the 
Bhrgus (martial priests, who presently attached themselves to 
Yayati and Ms descendants, specially the Yadavas) : Sukra- 
Usa-nas, Yayati' s father-in-law, married ' pitr-kanya 5 Go (or 
Ga). 7 This throws some light on the Kaca-Devayani story, 
where Kaca refuses to accept her as wife, as she being Ms 
teacher's daughter was ' equal to his sister/ but Devayan! 
insists (cf. Yami's insistence) and finally curses him for 
refusing her. 8 Devayani naturally regarded the excuse as a 
lame one, her father having married & sister (who was his 

1 Cf. n. 1 and 3, p. 118. 

2 Concerned mainly with the Aila and Aiksvaka kings, and closely 

connected priestly families like Bhrgus and Vasisthas. 

3 Vayu : II, 93, 12 3 Brahma : 12, 1; Hariv. : 30, 1599/Matsya : 15, 23; 

Linga : I, 66, 60-' 1 ; Kurma : I, 22, 5. 

4 Ma'tsya : 14, 1 ff. ; Brahmanola : III, 10, 54 f. 

5 Fable adds that the * pitrs * cursed her for this choice to be born 

again of Amavasu or Yasu (Caidya) as Satyavatl (Kali, etc.).. 
and the ' tithi * of the evil choice became Amavasya.* Such 
fables were obviously due to misinterpretation of ' pitr-kanya/ 
and in this case the starting points of the fable may have been 
the common royal name Amavasn (or Vasu), the Puranic saying 
that the Vasus were Pitrs (e-g-, Mai/sy a : 19, 5), and" the con- 
nection between e Amavasya ' and ' Kali-* It seems the fable 
about * Amavasya ' arose out of Acchoda' s appellation ' Amava- 
savi/ which again came to be confused with Vasavi (Satyavatl) ; 
probably Satyavatfs being ' pnnarbhii ' has also led to the story 
of the second birth of Acchoda; cf- similar confusion re Ajaml- 
4ha*s punarbhu wife, from which has originated the fable of 
Ajamidha's 2 births; cf. Vayu: 99, 206-9; Matsya : 50, 17-19; 
where ' punarbhave ' and ' putrabhaye * are apparently corrupt 
readings for ' punarbhava * (bhuvi, etc.) ; cf. Ugrayuxiha's 
would-be ' punarbhu ' wife Satyavati, in the same Pancaia line. 

6 As the Yayati story shows, besides other geographical references (re 

Vraparvan, Nahusa, etc.). 

1 Matsya: 15, 15; Brahmanda : III, 1, 74-77. (Sukra is Jhere said ^ to 
have been daughter *s son of Hiranya-kasipUj who-se Bister Simhika 
married a half-brother), * G'o ' was not a Tare name ; cf - 
Kakutstha's daughter Go, whom Tati married in the nest genera- 
tion; ana Suka's sister-wife Pivarl, also called Go. 
. 145 (Sambhavap. : Kaca) : I, 76-77. 

( 120 ) 

'dayita' wife) 1 ; her elder sister Devi married one 'Vara^a' 2 
and Eavi's immediate descendants ("sons") were called 
Varunas 3 ; so Devi may have married a brother or a first cousin, 
as gukra-Usanas was ' Eavya,' or KavIJkimself, according 

to one version. 4 Kaca himself, being an Angirasa, had little 
moral ground to refuse; for among the Angirasas, Samyu s 
second son Bharata married his three sisters, 5 and there were 
other incestuous marriages in the Angirasa- group. 6 As for 
marrying a preceptor's daughter, it is not very likely that 
custom was much stricter in Kaca's days, when so late as one 
or two generations after the Bharata war 7 a favourite resident 
pupil could be made the preceptor's son-in-law, 8 and even be 
asked (or allowed) to "beget children on his wife. 9 ^ Kaca's 
attitude therefore has no bearing on ' sister-marriage J in that 
age, but is an obvious case of political prudence, 10 ; : ust as the 
subsequent marriage of Devayam had an admittedly political 
significance. 11 

For about ten steps after this we lose sight of sister- 
marriages; then we get two very probable instances in the 
Aiia as well as in the Aiksvaka line, in the latter apparently 
for the first time since the semi-legendary Vaivasvatas, Yama 
and Mann. la each case the texts are muddled in 
the extreme, and obviously the different readings are futile 
attempts to rectify something that was ill -under stood or was 
considered improper and damaging; the motive was quite a. 
natural one, as in both cases the reference is to the marriage- 

1 Mafcsya : 15, 15? Devayani was Sakra's daughter by another wife, a 

daughter of an ' Indra,* _who may bs Raji who had become 
1 India ' In ids day ; cf . Apnavan, another Bhrgu of this time. 
marrying Euci, daughter of Nahusa, who also had become aa 
( ladra ' like his younger brother ; or Devayani's mother may 
hare been a daughter of Xahusa-' Indra ' himself; in any case 
her marriage with Yayati would be a consanguinous one. 

2 Mbh. 124 (Am&vat .) : I, 66, 2616. 

3 Mbh. 747, b. (Suvarnoip .) : XIII, 85, 4149. 

4 In Mbh, Sukra=Kavi; or Kaviputra som-etimes ; cf. Sorenaen : Index : 

p, 405. 

5 Mbh. 490 (Angirasa) : III, 219, 14135-37; thongh the account as a 

whole is mixed up with mythology, that does nob diminish the 
value of the detail quoted. (An Angirasa Samyu was somewhat 
earlier than* the historical Bharadvaja-Angirasa whose chronolo- 
gical position is fixed by synchronisms.) 
E.g., a daughter becoming a married wife : Mbh. 5 490 (Angirasa) ; 

m, 219. 

7 I.e., in Uddalaka-Arnni's time. 

8 Kahoda married Uddalaka's daughter Sujata. 

9 Svetaketa was so begotten on Uddalaka's wife ; cf . also the Vedic 

custom of transferring a widow to her deceased husband's pupil. 

10 The Angiras and f * Devas ' were at war with the Bhrgus and 

c Asoras * or Vrsaparvites, and Kaca-Angirasa's mission was to 
cheai the latter. 

11 As the Mbh. states, in reply to the question * fof/w I>evayani came 

to be Yayati' s wife/ that both Usanas and Vrsaparvan courted 
Yayati and sought his alliance. 

( 121 ) 

relations of the imrneliate progenitors or successors 1 of famous 

Aila and Aiksvaka kings. The Aila instance is further 
entangled in confusion, as there seems to ha-Ye been an 
irregular succession after Matinara, 2 and a gap 3 in the dynasty 
soon after this point, 4 as a result of the Hadhaya (Tadava) 
expansion and raids 5 (the great historical event of these 
times) . 

Of the texts that give an account of the Paurava King 
Matinara's descendants down to Dusmanta-Ailina (the reviver 
of the line), those of the Brahma and Harivamsa appear in 
this case to be the best 6 ; Vayu is here most corrupt, 7 and 
cannot be checked by the corresponding Brahmanda text 
which is lost ; the Matsya and the Mahabharata 8 have loosely 
followed and confused the two source-texts of Vayu and 
Brahma-Harivamsa, while the Visnu and the Agni 9 give very- 
brief and unsatisfactory summaries of these respectively. By 
collating these two latter texts first, and then that of Vayu 
with it, a proto-text may be approximately drawn up, 
specially as the source of the Vayu in this passage* seems to 
have been the same in spite of various corrupt readings. 
According to this collated text, 10 tl From Matinara, by 

1 Viz., Matinara and Dusyanta (Bharata's father), Alias; Prasenajit, 

Yuvanasva, Mandhatr, Purukutsa, Aiksvakas. 

2 Vide infra. 

3 This must be admitted partly on the strength of synchronisms , and 

partly because the undoubted Haihaya raids and supremacy im- 
plies prostration of the kingdoms of Madhyadesa for the" time 
being 3 so also, Kanyakubja, Kasi and Ayodhya are known to 
have fallen. 

4 I.e., between and Ailina-Dusmanta. 

5 From, Sasabindu son of Gitaratha and Mahismant son of Sahanja, 

to Jyamag-ha and Durjaya and Supratika (an interval of between 
13 to 20 steps), 
5 Br. 13, 51-55; Hariv. 32, 1714-1721. 

7 Vayu : 99, 121-133. 

8 Mat. 49, 710; Mbh. 1. 94, 3704 ff. 

9 Vis. IV, 19, 2; Ag. 277, 4b-6a. 
10 Collated prato-text : 

Matinarat Sarasvatyams trayo' jayanta dhannikab/Tamsur adyo' 
pra'ciratho Dhruvas capratimadyutih/sarve veda-vidas tatra brah- 
manah satyavadinah*/GaurI kanya ca vikhyata Mandfaatw 
janani tatba/ (putro* pratirathasyasit Kanvah sa nabhavau nrpah/ 
Medhatithih sutas tasya tasmat Kanvo'bhavad dyi|ah)** 
Ilina nama yasyasit kanya vai janamejaya 
(Or Ilina nama casyaslt kanya vai janamejaya. 
Or Ilina nu yarn! syasit kanya yajanayat sutan. 
Or Hinanupama tvasit kanya yajanayat sutan. 
Or Ilina tu pitur asit kanya sajanayat sutan. 
Or Ilina Matinarasya kanya sajanayat sutan. 

brahma-vadiny adhistri ca Tamsus tarn abhyagacchafca*/Tamso^ 
Surodho rajarsir Bharmanetro pratapavan/bralnna-vad'i 
parakrantas tasya bharyOpadanavi*/ Upadanavi sutam lebhe 
caturas ^vAilinatmajan/Dusmantam atha Susmantam Praviram 
Anagham tatha. 

* In Brahma and Hariv. texts only. 
** May or may not be spurious. 

( 122 ) 

(S Sarasvati 3 three virtuous sons were born, viz., Tamsti, the 
"eldest Apratiratha, and Dhruva, all of whom were truthful 
ki Brahmaiis learned In the Veda; and (he had) a famed 
'"' daughter, Gauri, the mother of Mandhatr " [here occur two 
lines (with variants, in all the three tests), which may be 
spurious, and wrongly inserted here 1 owing to a probable con- 
fusion between two Kanvas ; but as it stands in the collated 
test, it need not be so taken, for it rather explains what 
follows] ; " Apratiratha's son was Eanva who did not become 
" king; hence his son Medhatithi-Kanva became a ' dvijV " ; 
"blithe 31 (either Matinara, if the intervening passage is 
spurious, or Apratiratha, if it is an integral part, though even 
then " he " may well refer to Matinara, as the text is about 
him, and these two lines are by way of explanation only), 
" (but he) had another daughter named J Ilina/ a c brahma- 
"vadini* superior woman, whom Tamsu married, and who 
"gave birth to sons (i.e., heirs of the dynasty). In Tamsu's 
"line (were) Surodh'a, the rajarsi Dharmanetra, etc.* 5 Here it 
seems clear that Matinara had three sons and two famous 
daughters, and of these a younger son Tamsu married his 
influential sister Ilina, through whom the Paurava line was 
continued; if however the doubtful couplet is included, 
another possibility arises, that Ilina, instead of being the sister 
of Tamsu may have been his niece; in any case the eldest 
son Apratiratha's line was displaced by a younger branch 
strengthened by a consanguinous marriage. 2 

The Aiksvaka case is somewhat simpler. In each of the 

five tests. 3 collated Eere, the outline genealogy is quite clear : 
Samhatasva, the 4th predecessor of Mandhatr, had two sons, 
Ex^asva and Aksayasva, between whom and Prasenajit in the 
next step is placed Haimavati-Drsadvati, a ' famous lady/ 
the ' wife ' and the ' daughter ' of some of the persons named 
before her : while repeating this outline list, all the texts have 
evidently tried to gloss over some unacceptable feature in the 
relationship of this lady which is left vague. 4 On. collation, 5 

1 Of. Pargiter : AIHT, pp. 225-'28. 

2 The disqualification is apparently due to adoption of ' brahman '-hood 5 

* Eaava ' might also refer to the blindness of the heir-apparent. 

3 Vayn : 88, 63-64- and Brahmanda : III, 63, 65-66, forming one test ; 

Ha,riv. 12, 7G8-'tQ; Brahma : 7, 891!; and Siva : VII, 60, 72-'74, 
forming another. 

4 So also some other passages omit all details .regarding H&imavati : 

Mateya : 12, 33-34 ; Hariv. IV, 2, 13. 
i Colla-tfcd proto-test : 

Samhatasvo Nikumbhasya suto rana-visaradah 
Akayasva-Krsasvan tu Samhatasva-sutav nbhau 
(a) tayoh patnl Haimavati' sa-maturi, DrsadvatL 

or tayot patnl Haimavati sammata tu Dradvati. 
of tayot patmi Haimavati satam matad Dr^a^lvati. 
of tayo^ patui Haimavati tasya kanya \ 

pitr-kanya f 
vikhyata trin Ioken putraa casyah Prasenajit 

( 123 ) 

however, it becomes clear that the famous 
Drsadvati was a daughter of Samhatasva, and k< in accord- 
ance with authoritative sanction " was also the wife of both 
his sons, Krsasva and Aksayasva, so that Prasenajit was Jier 
son. Here, then, is & case of sister-marriage combined with 
polyandry : as Prasenajit was the grandfather of the famous, it was natural that this questionable feature of the 
original s vamsa-sloka J was sought to be buried beneath 
diverse guess readings. It is to be noted that these two Aila 
and Aiksvaka sister-marriages occurred in the same period 
(the latter being the earlier case). 1 

After two important royal marriages with the Paurava 
and the Yadava dynasties, 2 Samhatasva's line shows another 
instance of aster-marriage. Mandhatr's son Purututsa 
married his ' pitr-kanya/ 3Sfarmada, 3 who was later on, like 
so many other women of traditional history with names of 
rivers, 4 fancifully identified with the E. Narmada, 5 but is 
simply a princess in all Puranic genealogies. 6 In this case, 
again, there is probably a ' double * sister-marriage, a com- 
bination with polyandry, as in the case of Haimavati 4 steps 
above : the Brahma^da test 7 gives the sequence Mandhatr 
Ambarlsa (taking the second brother of the lists) = Narmada 

Yiivanasva Anaranya, etc., instead of the usual sequence 

Mandhatr Purukutsa = Narmada Trasadasyu Anaranya, 

etc., thus deriving the successors of both Purukutsa and Arnba- 
risa from the same sister-wife. As is to be expected, the various 
texts and readings at this point show signs of omissions and 

or (b) tasya Haimavati kanya satam matad Drsadvati 
(a line '?rob. lost; her) 

vikhyata trisu lo^esu putras casyah Prasenajit. 
or (c) tasya Haimavati kanya tayoh patni Br^advati 

vikhyata hi satam matat putras casya^i Prasenajit. 

1 For Prasenajit's sou married Matinara's other daughter Ganri, an 

alliance that forms one of the bed-rocks of Poraiiic chronology. 

2 Viz., YuvanasVa= Garni, and Mandhatr Yindumati. 

3 In the * pitr-vamsa ' sections of most Puranas (Matsya : 15, 25, etc. ; 

Br&hmanda : III, 10, 98.) 

4 E.g., Tapati; Kaveri; Kausiki, etc. Of. the numerous _ stories (in 

Pur.) of princesses being cursed and converted into rivers. 

5 As in Mbh. XV, 20, 549-'50. 

6 Vayu : 88, 74; Brahmangla : III, 63, 73; Brahma: 7, 95-'6; Hanv. 

12, 714- J 5. (Yinn : W, 3, 6-12, gives an account of how the 
Nagas (of the S.W. seaboard, from the contest) solicited Kar- 
niada to obtain for them the aid of Purukutsa against invaders, 
and she accordingly led him forfch into the Naga conntary in a 
victorious campaign (N.B. Mandhatr was already in fche S.W.); 
the N%as blessed her : ' there shall be^ no breaking off of thy 
offspring by Purukutsa * ; it seems possible, therefore, that the 
R. Narmada derived its nam* from the \ savionress ' Narmadl.) 
7 Brahman^a : III, 63, 72 ff. 

( 124 ) 

alterations; a comparison of these suggests a collated text, 1 
according to which,' 6 Of the three sons, of Mandhatr, 
ik Ambarlsa's heir was 'another' Yuvanasva begotten on 
4k aS T armad^; his son was Harita, from -whom were descended 
''the Hari(i)tas, military brahmans ; while Purukutsa's heir was 
"the famous Trasadasyu, begotten 'subsequently' on Narmada-, 
** and regarded as his"* own ' son : his own son was Anaranya, 
" etc." Evidently N armada was the wife of both the brothers, 
either at the same time, or by re-marriage (or * niyoga'). 
The Egredic version of Purukutsa's story, therefore, seems 
to embody a dynastic fact, 2 viz., that after Purukutsa's death 
or captivity, his queen (herself of the .same royal blood) 
obtained a" son for his race, and according to the Puranic 
indications, quite normally by her ' husband's ' brother, in 
this case also her own brother. 

The next group of instances of sister-marriage occur very 
much later 3 (21 steps below, according to one version, or 
37 steps below, according to another) ; and these cases belong 

to the Aiksvaka line again. According to the Matsya version 4 

Collated profco-text : 

(A) Purukutsam Ambarisam Mucukundam ca vis'rutam 
Ambariasya dayado Ynvanasvo'parati smrtah 
Narmadayam. samutpannali sammatayam tadatmajah 
(or Narmadayain samutpannah. sambhutas tasya catmajah. 
Halite Yuvknasvasya Haritah surayah smrtah 
ete hy Angirasa3a paksah ksatropet-ah dvijatayah 
Pumkntsasya dayadas Trasadasyur mahayasah 
Narmadayani athofcpanna^ sammatas tasya catmajaljt 
(or Narmadayam athotpannah san-matad tasya catmajaJ.i 
or Narmadayam athotpannas tv Amba(u)msasya catrnajah) 

ambhuto'syatmajah. putro hy AD&ranyah pratapavan 
or (B) Piirukntsam Ambarisam Mucukundam ca visrutam 

Narmadayam samutpannas tesazn sambhuya catmaja^i 
Ambarisasya dayado YuvaBasvo'parah smrtali 

Harito Yuvanasvasya d vi ja* ayah 

Ptimkntsasya pratapavan. 

TMs would make ihe hypothesis of 2 Purnkutsas and Trasadasyus 
largely .unnecessary ; ! Durgaha ' and * Giriksit * offer no real 
difficulty, as these are simply obvious epithets of an unapproach- 
able conqueror of the hilly S. W., whither Mandhatr was led by 
his wars and Yadava marriage; cf. * Trasadasyu/ an epithet 
derived from similar circumstances, used as a name. 
But a few steps below, in the time of Hariscandra-Aiksvaka, there 
was apparently a dynastic: custom and a * rsi * practice, of sister- 
marriage (and other incestuous cornections) permitted_ for the 
sake of oSsping (vide c Puranic ' gatha quoted in Ait. Bra.). 
That Hariscandra was effectively advised with regard -to attain- 
ment of offspring by the rsis Narada and Parvata, occurs in the 
present iPuranic tests also ; but the recommendatioa of incest is 
not there, as in th older * gatha ' text. It would seem as if 
the original bardic account of Hariscandrafs life and times has 
been retouched in questionable 'details by subsequent brahmanisa- 


Matsya : 15, 1349, 

( 125 , 

it was Saga-ra's grandson Amsumat who married ' pitr-kanya * 
Y-asoda, who is further specified as ' daughter-in-law of 
Paneajana, mother of Dilipa, and grandmother of Bhagiratha.' 
But the Brahmanda 1 distinguishes the * pitr-kanya* s -born 
Dilipa from the Dilipa who was Bhagiratha's father 
though in the ' pitr-vamsa * accounts the two are often 
mixed up. The genealogies in several Puranas 2 make Dilipa- 
'Khatvanga the son of ' pitr-kanya ' Yasoda, making her the 
wife of Yisva-mahat and * daughter-in-law ' of Vrddhasarman, 3 
or wife of the latter.* It is possible that both statements are 
correct ; the recurrence of sister-marriages in the two dynasties 
is too apparent to make this unlikely, and such a statement 
about the descent of Bhagiratha who was subsequently made 
into a brahmanical hero, is in itself proof of its authenticity. 
Repetition of names, even of women, is not unusual in the 
dynastic lists 5 ; and both Amsumat and Yisva-mahat (-saha* 
may have married sisters named Yasoda and had sons called 
Dilipa , 6 quite a common princely name. 7 

Visa-saha's sister-marriage was not however an isolated 
instance. At the 5th or 4th step 8 in his line, the famous 
Da^aratha seems to have contracted such a marriage with 
'Eausalya ' who can only have been a sister or a first cousin 9 
(paternal uncle's daughter), probably the former, as the 
cumulative evidence suggests. 10 It is to be noted that a 
Kaiisalya in the genealogies always means <a daughter of the 
Kosala king (of Ayodhya), 11 and never wife of a Kosala king, 
pure and simple ; and appellations of .similar formation, else- 
where in traditional accounts, have invariably and precisely 
the same import. 12 This gives added significance to the alleged 
succession trouble amongst Dasaratha's sons by his several 
wives : the rights of * pure ' dynastic blood could not be finally 

* Brahmanda: III, 63, 166; 181-182; 10, 90 ff. 

3 E.g., Vayu: 88, 180-182; Brahmanda: III, 63, 166; 181-182; 10, 
90 ffj in both, the misreadings * putrikasya,' ' putrikasl/ ' putri- 
kasyam/ etc., are obvious tamperings- with ' pitr-kanya/ and may 
date from a time when the * pit^-vamaa * explanation had not yet 
been devised. 

3 The names are variously read. 

4 Linga : I, 66, 31. 

s This has led to fables about the same 'apsaras'es, Ghrtaci, etc., 
being mothers of different kings in 'the same dynasty. 

6 Unless the two Dilipas are identified, from the standpoint of dynastic 
synchronisms; this point however still reqnires clearing up. 

1 There was a Paurava Dilipa also, before Pratipa, besides these two- 

8 According as ( Dli'ghavahu ' is taken as a name or epithet. 

9 This wonld probably shock those wbo have imbibed in good faith 

the medieval Eamayanic tradition. 

10 See the preceding cases, and also infra. 

11 Except in thse very few cases where Kosalan titles were used by 

conquerors of Kosala. (vide infra). 
I? See infra, the case of PrsatI or Parsatj, 

( 126 ) 

suppressed. Thus it becomes clear that the later Eavya version 
of the Eamayana is wrong in its statements about Rama, 1 and 
the Buddhist reference 2 that makes Bama brother and husband 
of Sita is historically right, in view of all this collective 
evidence. The origin of the modified version discloses itself 
in Sita's appellation 'janaka-duhita J which need only be 
compared with the ( pitr-kanya ' of numerous other Instances ; 
the transition from the substantive * janaka ' in what was 
probably the old basic genealogical sloka, to the proper name 
4 Janaka/ was a very easy one, and had the merit of supplying 
a plausible and honourable connection for the subsequently 
deified tribal hero, while removing the objectionable feature 
smoothly. 3 

Por 27 steps a-fter this 4 no sister-marriages are indicated 
in the dynastic accounts, 5 Then we find several cases again , 

1 As in fact in many other genealogical particulars, as compared with 

the consensus of Puranic traditions. 

I Gowell : ^ Jat, IV, 78-82. "it is to be noted that an early Buddhist 
version would originate in Kosala itself, and as Buddha himself 
belonged to the Kcsala dynasty (though probably a branch one), 
there can have been no motive of disparagement in such a state- 
.ment; besides sister-marriages and first-cousin marriages were 
not unknown in early Buddhistic period. The Jataka tradition 
indeed is base I on the very early Puranic, and it is quite likely 
that some real pieces of historical fact have been better preserved 
here than in later Bramanical works like the Eamayana, having 
been taken out of the earlier 'Parana' (9fch Cent. B.C.), within 
3 -or 4 cenfcnries of its collation. (N.B. The Kasmirian version of 
Sita's descent is a confusion between several popular cycles of 
stories connected with Havana, and cannot be regarded as being 
drawn from authentic Puranic tradition.) 

I The nature of the transition is well illustrated by a popular stanza 
cf an apparently unknown medieval Bengali ' Kavi ' (ex tempore 
epic and puranic dramatiser), which is intended to serve as an 
encomium as well as a denunciation on Rama at the same time : 
< Janama tomara ativipule/Bhuvana-vidita Ajera kule/ r Janaka 
dunita vivaha kari/Tahate bhasale yas'era tan." Evidently here 
is a trace of the earlier Buddhistic tradition (which lingered 
longest in Bengal). Many of the statements of Bamayana will 
have to be examined in the light of Puranic traditions and 
historical probabilities suggested by these latter : e.g., in the 
process of modernization and rounding angles, Siradhvaja may 
have been hit upon as a suitable c janaka ' for the ' janaka- 
duhita,' because of the connection between * sita * and c sira * ; 
f sayonija Sita ' of original tests may have been made into 
mythical c ayonija * Sita/ etc. ; one basis of identification of Sita 
with Janaka's dtr. was probably the story of VedavaAi, dtr. of 
Kmsadhvaja (of Mithila apparently), outraged by a ' Eavana ' : 
Bam. VH, 17. 

^4 The period may Ultimately proye shorter, when all "the synchronisms 
have been, more thoroughly examined; the present estimate is 
based on the taking of the* * solar * lines as the standard, and so 
there is room for corrections, 

5 Except another instance (noted infra) among the Ycdavas of Mathura- 
Sraena in the generation next to Eama. apparently under 
Koialan infiuenoe. 

( 127 } 

only one or two steps above the Bharata war. The Vafistha 
KrsnaJDvaipayana-Vyasa's eon Suia (the hero of many edify- 
ing brahxnanical didactic tales) married 4 pltr-kanya ' Pivari 1 
(who was anxious to obtain a worthy husband), just as, 
generations ago, the great Bhargava brahman, Sukra-Usanas, 
married a- sister. Though comparatively fewer instances of 
sister-marriages are recorded of brahman families, this is no 
indication of their rarity among- them, 2 but is merely due to 

* r ' 

the fact that it was only in exceptional cases of intimate 
contact with important ruling princes (like that of Sukra with 
Yayati and Vyasa with the Kauravas, etc.), on which much 
of traditional history turned, that such 'details about priestly 
marriage-relations were recorded ; for as a rule the brahman 
families kept no genealogies, 3 and whatever traditions are 
found about their sex-relations, show that they were much more 

3 t/ 

unfettered and loose in these, than the ruling nobility. 4 The 
other two instances somewhat less definite, are amongst these 
latter, Paficalas and Yadavas, in the same generation as Suka 
and Pivari (or Krtvi) . Drupada apparenty married his sister , 
and his sons and daughters, at least some of 'them, were by 
her, probably including Dhrstadyumna and Draupadl. A full 
account of Drupada's family is given in the Mahabharata, 5 
where it is stated that, intent on avenging Drona's insult, 

* In the * pitr-Yasns'as * generally ; Harlvamsa : 23, 1242- S 3 ? where she 
la caled Krtvi and a * pitr-kanya ' '(the "variation in the name 
but consistence in the epithet showing that it is a ( sister ' who 
is referred to) ; also called Go : Matsya : 15, 5-10 (where her 
daughter is called Krtv! and mother of Brahmadatta ; but Brahma- 
datta's maternal grandfather Suka must be a different person 
from Vyasa's son, though as Vyasa is said to have begotten 
Suka on a * guki * called Ghrtaci or Arani, the same family 
may be indicated by both references). It is quite likely that 
the wives of Jaigisavya (85) and Kasyapa Asita (91) , Ekapatala 
and Ekaparna, were the daughters of an actual brother-sister 
marriage, of Mena and Himavant, whoever they may have been. 
Of. Parigter :_ AIHT pp. 69-70; 192. Of. also n. 5 p. 75. 
g Cf. the definite Angirasa instance noted ante, and other indications 
dealt with there. In the mythological case of Skanda's children 
(the * grahas *), amongst whom the brothers are said to have 
been husbands to the sisters, (Mbh. 502 (Skandop ) :^ III. 230) 
it is admissible to recognize a reflection of primitive ' rsi * cnstoma 
or Atharvavedic (hence Bhrgvangirasa) ideas (it is interesting 
to compare Ev. X, 162 and VI, 55, 4), as the^ myth is a con- 
tinuation of the brahmanical story of the sis divorced wives of 
the Esis (Bhrgu, Angiras, etc.), to whom Skanda is affiliated, 
and as it falls properly within <the scope of that strongly brab- 
manical Veda. 

3 The so-called ' rsi-vamsas * being much later attempts at compEmg 

some account out of hearsay, achieve nothing else but a list 
of Gotras and a few Pravaras. jumbled up without historical 
order; probably these emulative attempts were due to the 
Furanas Jiaving subsequently passed into the custody of the 
brahmans after the Bharata battle. 

4 See instances infra. 

5 Mbh. 218 (Caitraratha. : Drau.-sambh.) : 1, 167. 

( 128 

and dissatisfied with his existing children, Drupada, for the 
sake of a suitable son, had a sacrifice performed by the 
Easyapas Yaja and Ms brother, who then summoned the 
Queen Prsati 1 (or Parsati) to the sacrifice, to ' ' accept the 
offspring/' but she raised some objections, whereupon 
Dhrstadymnna and Dr&upadi were miraculously produced 
without her, but were regarded as the Queen's own children. 
Putting aside the fable, it seems clear enough that Drupada' s 
queen was Prsati (or Parsati) , and she was, potentially, 
adoptively, or actually, mother of Draupadi and Dhrstadyumna, 
and she was also, the * mahisi ' (Drupada having apparently 
other wives), for she was summoned to the sacrifice. 2 Now 
Drupada himself was well-known as * Parsata/ being Prata/,s 
son 3 ; and ' Prsati ' (or Parsati ') can only mean daughter or 
grand-daughter of Prsata ; thus Draupadi herself is, in the same 
connexion, called c Parsati ' 4 (daughter of Parsata = Drupada) 
or ' Parsatasya svasa * 5 (sister of Parsata= Dhrstadyumna-). 
Hence Drupada-Parsata's wife Prsati (Parsati) was his sister. 
The other case is not equally clear : Satrarjit the Vrsni, a, near 
relative and a father-in-law of Krsna-Vasudeva, is said to 
have had ten sister-wives (or sisters as wives), who bore him 
a hundred children 6 ; they may have been his own sisters and 
half-sisters. But according to another less reliable version 7 
these ten wives were the daughters of the Kekaya king 8 ; while 

1 Pra,tl : Mbh. I, 6590 ; Parsati : Mbh. I, 6406. 

2 Of. * Kausalya * being the chief queen of Dasaratha; or ' pitr-kanya * 

Yaoda being the * srestha ' wife of Amsumat (Matsya : 15, 28} ; 
it is possible that the rank belonged to the sister-wife by cus- 
tomary right (cf. the ritual precedence of the sister over the 
wife in Ait. Bra.); the much discussed * Subhadrike Kampila- 
vasini. etc./ may after all refer to a Kampila princess of blood, 
the sister-wife and ' mahisi ' of the Kampila king (it is well 
known that Yv. ceremonials often refer to the Kuril and 
Pancala courts). 

3 In all Epic and Puranic genealogical accounts; cf. his several appel- 

lations derived from Prsata. 

4 Mbh. I, 6434; 7326; HI, 215; V, 55205 5565. Prsata's predecessor 

(interval uncertain) Somaka's chief queen was also a ( Parsati ' ; 
this implies that -there was an earlier Prsata before Somaka who 
too married a sister; in -that case this instance of sister-marriage 
would have to be placed shortly after Bama-Dasarathi and 
Sattvata's cases (vide infra). 

5 Mbh. II, 2349. 

6 * Dasa-svasrbhyo bharyabhyati Satru(a)jittah satam sutah * : Vayu : 

96, 53. There are a number of variant readings, all of which 
are clearly tamperings that have nevertheless failed to obscure 
the origisia.1 ' svasr ' and ' bharya. 1 

T Matsya : 45, 1719. 

8 In that case they would not be * svasarab. * proper, but cousins of 
Satrapt. his mother (or a near ancestress) being a Madri (Matsya : 
45, 1 fE; Brahmanda: III. 71, 18 ff)=3Caikayi; (Madia, Kekaya 
and Vailika are often indifferently used in <tfce genealogies; but 
these local particulars are unreliable in the case of the ill-kept 
Tadava onea) t 

( 129 ) 

yet another version omits all details 1 and notes only the ten 
wives and a hundred children, evidently because something 
was felt to be here, in the line of the deified hero 

a/ J 

Krsna's father-in-law. A collation of all the modified and 
senselessly corrupted texts, 2 however, makes Satrajit's poiy 
gamons sister-marriage obvious. 3 It is noteworthy that 
k sisters as wives ' without any distinct possessive reference 
occurs in another ease in the same family, where Bhajamana,, 
a son of Satvata (from whom Satrajit was also descended, 
and who himself apparently contracted a sister-marriage), 4 is 
stated to have married a Srnjayi/ whose son Vahya(ka) 
married the two daughters of * Srnjaya ' (or probably the same 
Srnja-yi '), being * bhaginyau s (sisters), and begat children 
on the ' arya(a)-bhagini * (elder sister). 5 Here k bhagini ' 
might refer either to the two wives as each other's sisters, or 
to them as own (or step-) sisters of their husband, while in 
any case they were his ' cousin- sisters J (also called 
* bhagini 's) 6 ; this ignoring of a sure confusion shows that 
' full s sister-marriages were also recognized by these Tadava 
genealogies, 7 even if such a marriage may not be clearly 
indicated in this particular case. A collation of the various 
tests, however, leaves little doubt on this point. 8 This 
probability increases when we find the above-mentioned Satvata 

1 Hariv. 39, 2076 ; Brahma : 15, 45 ; these are of course emended 

versions with a late Krsna-it-e bias. 

2 Two source-texts may be distinguished here : (1) Vayu : * Das'a- 

svasrbhyo bharyabhyah Satrajittah satam sutah * ; (2) Brahma : 

Hariv : Brahmanda : 

Dasa-svasrsu Satrajid-bharyasv a-san '] 

'? satam satah ' ( f tvasan s 

c, * -.- A j' ,,_.... li being an obvious emen- 

Satrajito dasa-svasr-bharyas tasain J -dation for * svasr/) 

3 In the same family and generation Jayanta is said to have married 

Jayanti, whose son was Subha. (Padma : V, 13, 99-100; for the 

names cf. Ahuka and his sister .iiuM in the same group); this 
too would seem to be a case of sister-marriage. 

4 Vide infra. 

5 Vayu: 96, 2-6 j Brahmanda: III, 71, 3-6 ; Hariv. 33, 1999-2QG3; 

Brahma : 15, 50-34 ; Matsya : 44 ? 47-50. 

6 So also, Duiisala is ' bhagini * of the Pandavas in the Epic. 

7 Which do not lack instances of other varieties of consanguinous and 

incestuous marriages. 

8 The Hariv. and Brahma test is evidently drawn up so as 'to evade 

the troublesome points. TJte Vayu, Brahmanda and Matsya 
texts with their variant may be thua collated : 

. f VahyakascO pavahyakah 

"Bhajamanasya Smjayyam ^ Vahyakayam ca Vahyakah 

" Srnjayyasca J S1lte dve tu Vahyakas te udavahat 

'* Srnjayasya I 

C dve susuvate 1 ,_ -,- 
- tasya bharye bhagmyau } ^ prasuyitam f satan vahim 

ye Vahya-d^rya-SrnjayyamBhajamanad vijajfiire...-jAyutajit, etc.). 
Vahyad anya-bha_giiiyain ye Bhajamanad vijajflire...... 

am Devavrdho raja ? etc.)." 



( 130 ) 

(son of Satvant and grandson of Jantu) marrying a ' SatvatI 
Kausalya ?1 (or perhaps better, a ' SatvatI' and a ' Eatiynlya'). 
This Jantu 2 married an Aiksvaki (Kausalya) k ; their son Satvant 
also 2 married a K&usalya, evidently a ' cousin-sister ' ; and 
their son Satvata, again, married * SatvatI Kausalya.' Here 
it is clear that this ' SatvatI ' can only have been Satvata' s 
sister 3 ; and if she is the same person as ' Kausalya?,' then 
this latter appellation can be explained as loosely applied owing 
to her being descended from a number of ' Kausalya.' s married 
into the family every generation, 4 or by the fact of traditional 
history that Satvata had reconquered the Yadava possessions 
lost to Rama and established his dominion over ,a portion of 
the fallen Kosala kingdom, 5 so that the Kosalan titles could 
be used by his family 6 ; but a collation of the texts would 
suggest that two different wives of Satvata and their children 
have been confused, and that originally the son of one of them 
was distinguished as 'bhaginya,' i.e., ' sister-born. ' 7 In any 
case, Satvata contracted a sister-marriage; and this is signi- 
ficant in view of the fact that he is a younger contemporary 
of Eama-Dasarathi, 8 in whose family there were several sister- 
marriages in that period, 9 and with whose family that of 
Satvata had intermarried frequently. 10 

If the Bharatai battle is taken to have occurred in about 
950 B.C. roundly (a quite moderate and reasonable inference 
from the facts of traditional history) , these last instances of 
sistevmarriage would be assigned to cir. 1000 B.C., by 

which time almost all the Egvedic suktas had been composed 
and were awaiting final compilation. In the light of these 
facts, the references in the Egveda to sister-marriages become 
more intelligible, and their significance gains perceptibly. 

1 Vayu : 95, 47 ; 96, 34. 

2 Ma'tsya: 44, 45-47 ; Brahma: 15, 27-30 and Hariv. 37, 1994-2000 

make Satvant son of Hadhn (instead of Jantu), bufc retain the 
ALkvaki mother. 

3 O'f . " Prati ' and ' Kausalya ' above. 

4 So that she had almost as much of Kosala blood as Yadava. 

5 Hariv. 95 5242-8; along with Vayu: 88, 185-6; Brahmanda : II , 

63, 186-7; etc.; also cf. Hariv. 55, 3060-96. 

S Cf. the case of the Haihayas Bhadrasrenya, Supra-tika, etc the 
later case of the Kas'i princesses Amha, etc,, being called Kausalyas 
as weJl; there was an Anslnara King of Kasi; cf. also the 
converse case of Rohini-Panravi (w. of Vasudeva) who should 
nave been called Bohinl-Madri or -Vahliki. 

f The best collation would be : 

Devavrdham nrpam 

bhaginyam, Bhajamanam ca Kausalya suuve sutam.*" 
(It w possible however to read ' Kaufclyan/ taking it an adi 

to sutan. ) J 

8 Vide m 5 above.. ' m 

9 Vide pp. 125-186 above. 

10 Vide n. 6 above. 

( 131 ) 

Bv. VI, 55, which shows ' rsl ' approval of Incestuous con- 
nexion with sister (and mother) , reflects the same state of 
custom and opinion as the old * gatha 7 (in Ait. Bra-..) refer- 
ring to the time of Hariscandra-AiksTaka, which agrees fully 
with the actual occurrence of sister -marriages amongst the 
Aiksvakas on either side of Hariscandra, and amongst ' rsi ' 
families, and is thus a very early reference. Bv. X, 162 
also is comparatively early, as it is an Atharva-vedic domestic 
charm, and as its evidence regarding the prevalence of such - 
connexions, at least in the brahman society contemplated by 
that Veda-, agrees_ with the early cases amongst the same 
groups (Bhrgus, Angirasas, etc.), as noticed in ' tradition/ 
But Bv. X, 10, which shows the rise of better opinion (and 
some conflict of opinion also), is clearly later 1 than those 
two ; hence it is best viewed as a * vakovakya ' or Puranic 
dialogue, of the character of a social drama on a small scale 2 
with a moral ; it is significant that the typical example selected 
for the moral dialogue belongs to the very earliest stage of the 
traditional dynastic history of the Aiksvakas (and Alias) : this 
indicates that the author knew Puranic traditions well, 3 and 
that the piece was probably intended for the reform of some 
Manva (or Aila) court and its attached priesthood 4 ; all this 
again, points 1o the time of its composition as being close to 
that of the bringing together of priestly and bardic lore in 
6 samhita's by Yyasa and his disciples. 

From this time (i.e. 1000950 B.C. downwards) , the 
Puranic tradition does not refer to any further sister-marriages. 
Though it notes some few details about subsequent dynasties 
for a century more down to cir. 850 B.C., for the succeeding 
period (850 to Magadhan ascendancy) it gives only the bare 
political facts and Gists of kings, without personal details ; yet 
there must have existed & mass of traditional history for these 
times, of which the stories about the Mngs contemporary with 
and preceding Buddha are surviving traces. Then in the 
early Buddhistic texts, which though fixed and canonized 
much later, can very well be taken as evidence for the 6th 
and 7th cents. B*C. 5 we get once again (some references to 
sister-marriage (along with other primitive forms) . 

An _ important question is raised here: Is this recrude- 
scence in the Buddhistic literature only similar to what the 

1 Ihis is^ also implied by its unknown authorship and subsequent 

ascription to the very persons who form its subject-matter. 

2 Of. the similar character of ' Pururayas-Urvasi ', and other pieces. 

3 Just as about 250 years later, the Aitareya made use of an earlier 

Puranic compilation (probably the one of 9fch cent. B.C.). 

4 So also, the reverse teaching of the (Puranic) c gatha ' in the Aitareya 

is for the benefit of the Manva King Hariscandra, put in the 
month of ris patronised by him. 

previous dynastic history in the Puranas reveal, or is it the 
effect of some external influence and change in social elements? 

The Interval between the last Puranic-Vedic instances 
and the Buddhistic references is not too long to make the 
first view improbable, when similar previous intervals are 
compared. In fact these intervals of no information are no 
proofs against such practices, and the recrudescences may as 
well be taken as marking a continuity in dynastic or priestly 
custom. If the Parana had not been closed, the continuity 
would in all probability have been well illustrated : it is indeed 
indicated by the fragments of non-Puranised tradition 
embodied in the Buddhistic texts. These Buddhistic texts 
are not all ' Buddhistic ' : among them are echoes from the 
older Puranic traditions regarding the pre-Bharata times, such 
as Eama's marriage with his sister Sita, 1 or Krsi^a's twin 
brother's marriage with his mother's daughter by" her second 
husband 2 ; or again, allusion is made to dynastic details at 
some stage or other in the post-Bharata and pre-Buddhistic 
period, such as the Kasi prince Udayabhadra's becoming 
the heir-apparent by Ms marriage with his half-sister Udaya- 
bhadra, who proved a most devoted wife 3 ; while another 
reference might belong to Buddha's own times, such as the 
proud admission^ of the Sakyas (a section of the Aiksvakas) 

1 Cowell : /atakas : IV, 79-82, etc.; vide also ante, re earlier aister- 
marriages m Rama's line (sp. pp. 125-126 and n. 2 & 3, p. 126). 

g Aec. to the Jataka version, Dranpadi and her brother were really 
children of the vanquished Kosala, King, their mother having 
been abducted and married by the victorious ' Kasi * King, during 
her pregnancy; -after the birth of the twins, the son was for 
safety brought up in secrecy away from the King's household, 
wnile tto daughter was recognized as his own; subsequently the 
boy fell in love with his mother's daughter by her second 
consort, and being caught in her company and recognized, was 
duly married .to Ms half -sister (vide Cowell ; Jatakas : V, 226, 
etc.). ; 

These dynastic details agree very well with those In the Epic and 
Puranas re the Pancala line : Drupada himself married a sister, 
and his ancestor Somaka did the same (vide ante) so it is quits 
likely that Dhrstadyumna also contracted a similar alliance, and 
the practice was in accordance with Pancala dynastic tradi- 
tion- (Of. also the * miraculous birth ' of Drau., and Dhrsta. 
in the Epic). v 

The selection was made after a good deal of search for a girl after 
the Wongs of Udayabhadia. The story makes them rather 
unwilling _ parties to the marriage at first ; but this is probably 
an addition, as the subsequent portion of it shows : after tho 
brother's death the sister continued to rule the country, and 
firm in chastity could not be seduced by others, as she longed 
for reunion with her lord and brother Udaya ; subsequently she 
abdicated, retired as a recluse, and " became the wife of Udaya 
again * ; " in fact she was Buddha's cousin-wife in a later birth." 
{Cowell ; Jatakas : IY, 67). 

( 133 ) 

that amongst them men ordinarily consort, wiffi their sisters. 1 
Hence a continuity of the custom seems to be clearly indicated 
during the interval in question. 

On the other hand, the Persian expansion into India from 
the first half of the 6th century B.C. onwards, makes it very 
prcbat le that kindred Iranian court- 4 fashions s2 were taken 
up in Indian aristocratic circles at that period or even some- 
what earlier. This does not imply anything like Spoonerian 
ZoroastrianisatioEu The Puranic tradition helps us in viewing 
the so-called Iranian influences in their proper perspective. It 
looks upon these Trans-Indus peoples of the far West and 
North- West, as being originally Aila (and partly Mksv&ka) 5 
communities, that migrated (or were pushed back) thither from 
Madhya-desa (along with' other offshoots to the S. W., etc.), 
at sundry times, but chiefly during the period from Yayati to 
Uslnara " (cir. 1900 1650' B.C. in Puranic computation). 
And throughout the traditional history of the pre-Bharata age 
they are never wholly lost sight of, at least the more easterly 

^ I v_? / *. 

sections of them, though often termed * barbarians/ etc. 
Indeed it seems very probable that the various 4 Harbarian ' 
inroads 4 from the N. W. and W. referred to in dynastic history, 
e.g., temp. Kusika, th Haihayas and Sagara (cir,. 1650 

* Dowel! : Jatakas : V ? 219. (In a tribal quarrel the Holiyas charged 
tie Sakyas with having this incestuous custom : tie Sakyas 
retorting, acknowledged it, saying that these sister-marrying 
Sakyas were mightier men than tie Koliyas.) 

2 Every student of history knows that many West Asiatic dynasties 
cherished the custom of sister-marriage, e.g., the Ptolemide and 
earlier Egyptian 'dynasties, the Achamenians and Sassanldes, etc. 

5 Thus a section of the Manvas is jsaid to have migrated beyond ^ the 
Punjab and become known as Sakasj and the Drahyu-ite sections 
of the Ailas beyond Gandhara came to be called Yavanas. 
[Q. Has the alternative name Dranghlana of Siestan and 
Arachosia (ace. to the Gks.) a connection with Drungio 
(Druhyu)?] [This Purardc tradition re migration of Manva and 
kindred tribes westward to the Punjab and adjacant countries 
early in the 2nd millennium or in the 3rd millennium B.C., seems 
to be substantially correct from the nature of the Harappa and 
Mahenjo Baro excavations of 1924]. 

4 Of Sakas, Yavanas, Kambojas, Paradas and Pahlavas, forming one 
erroup, and of Nagas, Abhiras, and Nisadas, forming another. 
The Sakas, etc.. who invaded Madhyadesa with the Haihayas, 
came from the highlands beyond the Seven Rivers : the first group 
wonld thus represent purer Alias, Airyas. or the Iranians proper) ; 
while the second group, coming by way of the sea, the lower Indus 
.and Sarasvati, up to Gnjrat, Bajputana and the Punjab, would 
consist of various races with non- Aryan affinities (probably partly 
Elamites, Snmerians and Semites of the Eed and Arabian sea- 
coasts, who had pre-historic connections with the Dravidians and 
the S. W. of India). The notion that every occurrence of 
Sakas and Yavanas refers to the Kushans and Bactrian Greeks, 
must be modified : these names quite naturally came to be 
applied to all invaders from the Western regions beyond th 
Indus, whither the original Sakas and Yavanas had migrated 
from T 

c 134 y 

1450 B.C.), or on the eve of and after the BEarata war (cir. 
1050 850 B.C.), were of these ousted members of the Indo- 
Aryan stock, in the tide of return. In the social history of 
the early Indian ruling classes and connected priesthoods,- 
therefore, the periodic reappearance of primitive types of sex- 
relations may have been, in some cases at least, due to strains 
of ' barbarian ' (W. Aila or Iranian) blood and practices 1 
from time to time ; and the Iranian expansion and influence 
of the 6th century B.C. would seem to have been merely a 
repetition of history. Hence few things absolutely foreign to 
Indian culture and traditions could have been introduced by it ; 
and if as a result of the Persian conquests there were any 
social changes, these would be mainly reversions to, or modifi- 
cations of those common features of" Indo-Iranian (i.e., early 
Alia) culture, which may have been retained longer, or special- 
ly developed, in the Iranian or (Druhyu-ite)" sections. It 
thus becomes intelligible how Persian influence in the early 
Buddhistic period could have led to a revival (however tem- 
porary or limited) of extreme consanguinous marriages. 
Similarly^ the ' sale of brides s and f exposure platforms ' 
at Taxila in the Persian period would be based on, and revivals 
of, the * asura ' custom of bride-selling as praticed (in the same 
area) by the Madras of tradition 2 * from time immemorial 1 / 
and the sporadic usage of exposing the dead as noticed in Vedic 
texts. All the so-called Persian features may be thus viewed 
and explained through ' tradition/ without any far-fetched 
theory. That the Ksatriyas of Kapila-vastu and Yaisuli were 
foreign races from Tibet or some other unknown land who 
developed a new and a crude type of religion and culture 
would be a supposition too fanciful and superficial to be enter- 
tained in this connection. The Puranic tradition knows the 
Salrras as a part of or offshoots from the old Aiksvaka race of 
Jiosaia, and VaiSali and Videha as continuing under the rule 
of cognate Manva families down to the close of the 5th cent 
-B.U ; wide all that is known about the early historv of 
Jaimsm and Buddhism show that they began as enlightened 
movements for betterment and reform in all 'directions, and 
arose from within the existing elements ; the only external 
influence that may be suggested to have worked, can be the 
rapid expansion of Persia at the expense of India, which 

is < said that consangmnous and 

,' cteriBtic of the < mfecchaa ' and 

Even if Sf T^l V P f C \r ? aisSciki ' ( Indo-Iranian dialect). 
this wlw 1 1 Ind^fldlmg u taken as a Mesopotamia* feature 
this would bejfchere as much because there were earlier West- 

^ tl L T 1 w 10 ? < thro ^ h sea-faring Nisadas who traded in 
girls at the Western port 9j and Nagas who Med at Taksa-sila 
oil the eye of and after ihe Bharata war), as on account 
migration of institutions within the hete^genous San 

( 135 ) 

must have given some sort of an impetus towards improving 
existing conditions. Buddhism or * Jainism in themselves, 
therefore, cannot be supposed to have introduced primitive or 
consanguinons types of marriage ; they were rather a source for 
purity and higher standards in sex-relations (as in many other 
lines of life and conduct) . The so-called high Hindu ethics 
and personal morality of subsequent periods, is very largely 
a Buddhistic achievement, a lasting reform and refinement, 
inherited bv later forms of Brahmanism. 

v * 

The above Buddhistic references are thus partly echoes 

from, and continuations of, the Pura-nic tradition, partly a 

reflection of lingering practices, and possibly in part 
indications of some Iranian influence (consisting in direct court 
examples and indirect preparation of an atmosphere for 
revivals of ancient and common Indo-Iranian (Aila) customs 
that were gradually falling into disuse in India after the 
Bharata war) . 

Incestuous marriages, however, must have continued far 
enough into the Buddhistic period to make it possible for 
the Indianised dynasty of Siam to have or retain a custom 
of sister-marriage by preference, even in later medieval times. 
Euling families and priesthoods intimately connected with 
them', have always lagged behind the line of popular progress 
in such points of culture; and in ancient history generally 
we find them sticking to obsolete and primitive customs : this 
is equally true of India. 1 The late and not uncommon per- 
formance 2 of the revoltingly primitive rite of the Asvamedha, 
in spite of early protests from Ksatriya kings (like 
Janamejaya-Pariksita IT, cir. 900 B.C.), 3 &n<i subsequently 
from the Jama-Buddhist reformers, shows the tenacity of old 
barbaric practices and their continuance even after a much 
higher level of culture was attained generally. And thus it 
must have been with sister-marriage and other crude tvpes of 

1 Where on either side of what may be called -the real * higher classes * 

have existed remnants of earlier stages of culture : with the 
ruling aristocracy and connected priesthood OB the one hand, and 
the gradually aryanised aborigines on the other. 

2 E.g. In the Snnga and Gnpta periods. 

3 Of. Hariv. 192. 11092 it ; 195-'6, 11236-'6& 

- 1 * * 



Quite In agreement with the Vedic evidence on the point, 
we find in the Epic-Puranie tradition a few plain instances 
and some indications, of incestuous connexions of this type, 
some of which might be called marriages. 1 

The mythological reference in the Egveda (X, 61) to the 
union of Prajapati with his daughter finds its counterpart in 
the Purana as well. 2 But whereas in the fortner the treat- 
ment and setting is cosmogonical and allegorical, 3 that in the 
latter is semi-historical ; and it would appear that the Vedic 
composer, Nabha(ga)-nedista Manava, 4 utilised an ancient 
and current tradition regarding the first origins of a previous 
ruling raoe, probably taken from the ' suta-magadha ' bardic 
accounts 5 of the Prthu-ite dynasties that flourished in N. B. 
India during the two centuries (or more) before the rise of 
the Aiksvakas and Ailas. According to these accounts, 6 the 
first famous chiefs in that earlier period, Priyavmta and 
Uttanapada, were sons of a ' Manu ' who was begotten 7 by 
Brahma ' ( = ' Prajapati,' 8 etc.) on his own daughter 
Satarupa whom he loved. 9 Sometimes it is explained how ho 

1 One instance is actually taken as a maniage in Mbh. | vide infra. 

2 In the accounts of ^ the origin of the First Dynasty of traditional 

history (in which Priyavrata, TJttanapada, Dhruva, Bharata, 
Vena and Prthu were the famous names, and which produced 
several * Manu's and * Prajapati's, and also the first ' Kinga * of 
India and their chroniclers and panegyrists, the Sutas and 

3 Quit in agreement with the usual want of rsi appreciation of his- 

torical traditions. 

4 A brother of Iksvaku, and progenitor of the Manvas of Yaisali', ^in 

the same region where the Prthu-ites once ruled, whose begin- 
nings ax placed by tradition in the 17th step (i.e. cir. 200 years) 
before Nabha(ga) nedista. 

5 It is to be noted here that ' Suta ' traditions were older than the 

Aila-Aiks.vaka period, dating from at least a century before them 
(i.e., the reign of Prthu). (The Puranas profess to give one 
ancient * Suta ' ballad, re Prttou's reign). 

6 I.e. the version of them preserved in the Puranic compilation of a 

later age. The interval between the final compilation of Puranic 
tradition (cir. 850 B.C.) and the 1st step of the Prthu-ite dynasty 
would be about 1400 years (=100 steps after Mami+17 steps 
before him). 

? With six others : Matsya, : 4, 2432. 

a Possibly the Vedic legendary version arose out of a confusion 
between Prajapati ibe divine creator and Prajapati the usiaal 
Puranic designation of early chieftains, whic& latter sens, evi- 
dently, is to be understood of Priyavrata's grandfather in the 
dynastic lists. 

9 Matsya : 3. 31-44 ; 

{ 137 

did not incur guilt by such a connexion, and stories are told 
of his curse on Kama, who became later on Pradytimna, and 
after the Bharata battle, son of the Yatsa King. 1 This sub- 
sequent explanation of the incest is paralleled by the 
Brahmana commentaries 2 on Ev. X, 61; and the basis of the 
appended fables was probably similar incests originally also 
recorded of Pradyunma and the Yatsa prince, 3 the case of a 
near ancestor of Pradynmna being actually on record, 4 
Another Puranic yerslon, 5 however, makes Satarupa the wife 
of Svayambhuva Mann* 6 instead of the daughter and wife of 
Ms father : this is either a subsequent improvement by one 
step on the older version, or might imply a- double incest 
involving another of the reverse order, which seems to be not 
altogether unknown to early legend and tradition. 7 A parallel 
is afforded in the case of Yivasant and Mann, both being 
called c Sraddhadeva/ while Sraddha is a daughter of the 
former and wife of the latter. 8 The Egvedic conception oi 
Pnsan as the * didhisu J of his mother, 9 and ' jara * of his 
sister, 10 and the * rsi * advice to Hariscan'dra (Manva) that 
the sister and the mother were permissible wives of one 
desiring offspring, 11 show that the two statements regarding 
Satarupa are not contradictory. 12 Tradition also supplies 
similar particulars ab'out the priestly groups : in an ingirasa 
genealogy (partly tinged with myths), ss the maiden Eohini, 
daughter of Hiranyakasipu s} is stated to have become " Ms 
"bharya * as a result, of * karma/ " 13 which agrees with the 

1 Matsya : 4, 132. 

2 Ait. Bra. HE, 33, 5; etc. 

3 This latter suggested instance would Indicate that aristocratic morals 

had not improved much, in the interval between the Bharata 
battle and Buddhism, (vide the sec. on sister-marriage). 
* That of Taittiri and his daughter; vide infra. 

5 Visnu : I, 7, 15-16 (where, as well as in Hariv., the sin is cleansed 

" "by Satarupa's penances cf . * aparfipam * in the corr. Matsya 
text) ; Hariv. 2, 54 fE. (prob. f patnim gatarupamayonijam ' in 
one of these verses is better read as * sayonijain '). 

6 Svayambhuva Mann is called an * Apava * in Hariv. op. cit. ; as 

* Apava f is a real clan name in Fur. tradition (applied to Hima- 
layan Vasisthas), it would seem he was a- historical^ person and 
not i a mere abstraction standing for the first origin of the 
Prthnite dynasty. 

7 Vide Infra. 

8 Vide ante, sec. re sister-marriage. 

9 Of. the legend in the Epic, of Mahadeva as Parvati s child on ner 

lap,: Mbh. XIII, 161. 

10 Vide ante, sec. re Vedic evidence on this subject. 

11 Vide n. 10 above. . , 

12 This is further supported by the fact that Satarapa is also called 

A.nantarupa or Anantaf !), and this name is given in some texts 
to Svayambhuva Mann's wife; cf. '...aparupam Ananta nama.., 
in the Matsya text. (A similar incest seems to be Deferred to 
In some .subsequent corrupt lines in the Hariv text Jcited above), 
about P?iyavrata and Kamya (nest generation) ; it is to be noted 
that Anna's sister-marriage also belongs to this dynasty.). 

13 Mbb. 490 (Angiraaa) : III ? 219-20, 141W, cf . Nil, commA 

( 138 ) 

fact that the ' Angirasa ' Veda also knows of actual cases 
like this. 1 In this connection the epithet ' Kanya-Bhartr ' 2 
appears significant, as applied to Sbanda in the brahmanical 
legend of Skanda's birth (of Atharranic character). 

But apart from very early or semi-legendary instances, 
tradition also notices much later and actual cases of incests of 
this type; and it is noteworthy that the two definite 
occurrences are ascribed to the Vrsni (Tadava) family, 
otherwise remarkable for laxity. Between 6 to 11 steps above 
Krsna in the Vrsni dynastic lists, was a musician king, 
Taittiri, who personally instructed his daughter in music, 
dancing, etc., and becoming enamoured of -her, begot a son 
IsTala on that daughter; hence Nala (who succeeded him) was 
nicknamed ' Nandanodara-dundubhi.' 4 These details are by 
no means fanciful, as the dynastic lists wherein they occur, 
are full of all sorts of natural personal details, 5 and kings 
instructing their daughters in music and dancing is quite a 
common thing in the dynastic accounts : thus in the same 
(S. W.) region and the same (Yadava) group of ruling 
families, Burjaya (the Haihaya) in an earlier generation taught 
these arts to his sons and daughters by a * gandharvi ' 
(i.e., a court dancing-girl) 6 ; a few generations below Nala, 

1 Vide ante, re Vedic evidence on this subject, 

2 Mbh. Ill, 14653. 

3 Nearer 6 than 11 (adopted by Pargiter), as all Puranas practically 

agree in the list from Kapotaroman to Kamsa. 

4 The full accounts are in Matsya : 44, 62 ff ; and Kurrna : I, 24, 49-54; 

other Puranas summarise,- some give only the nickname of Nala : 
often in corrupt unmeaning forms ( which makes it possible that 
Vasudeva's appellation ' Anaka-dundubhi ' is a modification of a 
nickname like ' Kanyaka-dundubhi,' and points to a repetition _ of 
Taittiri's case in the* family) ; and some simply give the succession 
list without any particulars; for fche former cf. Padma : V, 13. 
47-51; Visnu: IV, 14, 4; Brahmanda ; IIL 71, 117-119; 
Vayu: 96, 117; and for the latter, Hariv. 38, 2016-17, which 
being a specially Yadava chronicle omits Nala as well as his 
nickname. Nala also was musical; cf. Visnu above. The words 
'snsuta* and 'viloman. 1 in the Padma and Visnu list "seem to have 
been descriptive of Nala's parentage in the original verses, rather 
than separate proper names (i.e. =' svasutayam ' and * vilomaja *; 
cf. Matsya : 44, 63, where ' tanujali sarpo ' is obviously a corrupt 
reading for e tanuja-garbho(e).* Vide n. 2, p. 143. 

5 E.g., " Nala's son Punarvasu was born in the middle of the assembly 

at an Asvamedha," etc. Matsya: 44, 64-5; etc. 

ft Kurma : I, 23, 644. The daughters subsequently married 
* gandharvas * and the sons * gandharvis ' : a detail indicating 
that the lighter pursuits of the Yadava courts tended to produce 
a general laxity in their marriage relations. Cf. QPururavas 
marrying 'a * gandharvi % 'and their sons also doing tbe same, and 
associating with ' gandbarvas * and c gandliarvias ' together with 
their father: Kurma ; I, 23, 46: lor other refs. vide Pargiter: 
AIHT. p. 297, n. 5-8. 

( 139 ) 

theite is the well-known instance of the musical Revata 1 and 
his favourite daughter (about whose marriage there was some 
difficulty) 2 ; and in the next step there is a somewhat similar 
case of Arjuna instructing a friendly cousin's daughter Uttara- 
Vaarati (who evidently became enamoured of him). 3 The 
probable indication of a parallel to Taittiri's case in the fable 
about Pikdyumna's being * Kama ' by Brahma's corse, has 
already been noticed. Thus the early Yedic references 4 to 
actual father-daughter connexions are confirmed by the 
traditional. 5 

Pusan's position as ' didhisn s of his mother has, how- 
ever , no * specified ' parallels in tradition, except in the 
already noted mythical or semi-mythical statements about the 
two l Manns,' and the ' gatha ' allusion to the custom, in 
Hariscandra's time which, taken together, would suggest 
that amongst the Manvas and connected brahman families, 
there was a practice of transference of the father's wives to 
the son. But connexions between persons in ' similar * 
position are specified, and were probably frequent. A clear 
case is that of Samba, 6 son of Krsna, whose connexions with 
his step-mothers 7 are said to have brought Ersna's heavy 
curses 8 on him as well as the wives, the initiative in the 
affair apparently belonging to Samba's * mothers *f; so, also, 
when Satyabhama-Satrajiti seets from Draupadi the secret of 
her power over her five husbands, the latter warns her 
against talking or staying in private with her step-sons 
Pradyunina, Samba", etc. 9 With this may be compared the 
story of Arjuna and Urvasl. 10 

On a careful consideration of all the 'dynastic relation- 
ships 'described in the Epic, it becomes clear that the stories 
about the miraculous birth and marriage of the Pandavas are 
all late after-thoughts, only of value as showing that after all 
they were begotten by just the ordinary type of Epic Myoga 

1 The Saryatas (whose priests, too, were Bhrgus) became early affiliated 

to the Yadava-Haihayas, and became scattered in the S. W. 

2 In all Puranic accounts. 

3 Gf, Mbh.,, "Varataparvan. * 

4 Vide ante. 

5 To these may be added a tradition that c Havana * -would or did have 

his daughter by Mandodarl as consort (who was reborn as SitS). 

6 Yaraha : clxvli ; cf . Matsya : 70 ? 2 ff ; etc- So also his brother 

Pradynmna marries his foster-mother Mayavati, gambara's wife, 
ivho takes the initiative, and discloses to him that she had only 
nursed him, and did not bear him. (The Pnramc account tries 
to show that this doubly unseemly onion ^as justified.) 

7 ' Krsna's wives ' and ' Samba's mothers,' without any specification, 

8 Of "leprosy and prostitution, respectively. 

9 Mbh 51011 (Drau.-Satya.) : HE, 233-*35. 
10 Mbh. Ill, 45-46, 18121867, 

( 140 ) 

or license, and married according to a form not yet totally 
obsolete other cases of polyandry being known to the Epic 
and the Pnranas, and instances of the raising of offspring by 
relatives or outsiders, and of illegitimate natural sons, being 
quite common amongst the ruling and priestly classes of those 
times. 1 It is thus evident that the fables in the present version 
of the Epic and Puranas regarding the Pandavas, arose out 
of actual but (according to later views) discreditable relation- 
ships, and it may be possible still to discover traces of what 
the original facts were like, divested of fabulous garb. Leaving 
out further details on this point, 2 it may reasonably be taken 
to have been an * original ' fact of the Pandava history, that 
the person (called * Indra ' etc., in. the fables), who begot 
"Aijuna by 'niyoga,' received Arjuna in his court, 3 when he 
left the rest of the family to prepare for the battle, and mate- 
really helped 4 him with arms and training, and also entertained 
him right royally. The Arjuna-Urvasi episode comes in. here. 

Shorn of ' pantheonic,' legendary setting, the substance 

of it Is that one Urvasi, a chief dancing-girl attached to 
Arjuna's ' father's ' court and recognized as being in the 
status of his ' mother/ became enambured of !Arjuna (who 
was being instructed in music and dancing in her company) , s 
anfl, with his 'father's' consent, approached him; but 
she was refused by him on grounds of higher morality 6 (she 
being " guru-patm ') , though ,she pointed out that, in accord- 
ance with custom, all Arjuna's forefathers, the great Paurava 
princes, had accepted precisely similar invitations, without 
any guilt being attached to them. 7 There are indications 
that make it probable that the ' father ' of Arjuna was a 

1 Vide details in sees, re polyandry and ' niyoga.' 

2 Of. infra, setc. on * niyoga/ re KuntL 

3 At * Amaravati,' which may well have been a real city (of Central 

In ^ a ' vide n. 1, p. 141) so also there was a real Tripura and 
a Vaibhifaja in traditional history. The transference of 'the whole 
scene to Trans-Himalayas is evidently due to the c Indra ' fable, 

4 -** noticed later on, the three chief and original su-pporfers of the 

Kaunteyas are also very likely persons to have :>een their mo- 
gemtors by ' niyoga/ 

5 Note the specially Yadava and South-Western feature, and the 

parallels noted above. 

6 The Epic emphasises the 'great merit of this story of restraint ' 

on the part of a prince ; the parallel in the Puranas of the ' great 
merit that is d aimed for Arjuna's great-grandson Janamejaya 
ILL s story of opposing obscene ceremonials, is striking. This 
indicates that puritanic stands were exceptional, and laxity and 
tebanma were the general rule with the Yadava and Paurava 
ruling classes and their priesthoods, 

T Her curse oa Arjuna has a remarkable 'harem' tone, which is 
probably more than accidental. 

( 141 

I 7 adava prince related to Eiin&f s family, and he may have 
hfeen Purujit the Kuntibhoja 1 ; this would agree fully with 
what tradition says about the harem life of these 1' 
families, wherein such ' artiste '-concubines and las morals 
were a chief feature. 

These episodes of Samba and Arjuna point to an estab- 
lished dynastic custom, amongst the ladavas and Pauravas, 2 
of sons succeeding to the seraglios ( ' official ' or L non-omcial ' ) 
of their father very late medieval instances of which have 
been known in India as well as in other countries/ The 
arrangements which were made by Arjuna after the fateful 
slaugnter at Dvaravatr* make the probability surer. 'JLiie 
wives of the princes who had perished, were divided into 
three batches, and the three surviving young princes of Krsna, 
Satyaki and Krtavarman's direct lineage succeeded to them, 
and. were established along witu tliem in new principalities. 
So also Yicitravirya's wives are proposed by his mother to be 
transierred 5 as wives to BMsma, wno is requested to succeed 
him only in this case by exceptional circumstances the 
proposed successor is an elder brother. So, again s the palaces 
of Uuryodhana, etc.., are, after the battle, transferred to 
Yudhisthira's brothers, who spend the nights happily there 
the inmates of the palaces may have ben transferred too 
along with them. Such transfers would naturally involve 
incestuous connexions in the case of dire-ct lineal successions. 
This is illustrated by the definite statement in Vats. Ka. Sut. 
(referring to practices of the post-Mauryan or possibly a much 
earlier period) that the princes of Vidarbha (Yadsroa), in 
accordance 1 with ancient custom, freely consorted with all their 
father's wives, excepting their own mother. 7 The later Sutra 
dictum., therefore, that property in cattle, land and women, 
is not destroyed by changing hands, 8 is in part a laconic 
crystallization of much more ancient customs. 

1 Vide n. 1, p. 140. 

2 As apparently among the Manvas, vide p. 139 above; cf. Cowell : 

Jafc. VI, 133, for a Magadhan case, apparently of the Epic age, 
where Dirghavahu receives his father Arindaraa's 16,000 wives in 

3 E.g., the famous medieval case of the Bajpnt. princesses of Gruzrat 

(mother and daughter) being transferred to successive Delhi Em- 

4 Mbh. XVI, 7. 

5 Mbh. 168 (Bhis.-Saty.) : I, 103, 

6 Mbh. 637 (ESjadh. ) : XII, 44, 4147-'68. 

7 Vats. Ka. Sut. V, 6> 12. 

8 Gautama : XII, 39. 

( 142. ) 

The episode of Uttara' s marriage with Abhimanyu ,(in the 
Epic) can now be viewed in the light of these observations : 
As Aijuna taught Uttara music and dancing, the first thought 
that occurs to the court is that they should get married as a 
natural sequel 1 ; in fact Uttara is described as being clearly in 
love with Aijuna 2 , and she was a fully developed young woman * 
and no toying child 5 ; in spite of -,all this, she is married to 
Arjuna's son (barely 16), probably younger than herself. 4 
These details, therefore, are quite in keeping with the dynastic 
traits noted bv tradition. 


The Vedic evidence, considered by itself in a previous 
section, supplies no definite clue as to the nature and origin 
of the incestuous sex-relations there referred to. But the 
complementary evidence of traditional accounts helps in arriv- 
ing at some reasonable estimate. Taking the two together, 
it seems clear enough that these references fall mainly into 
two classes, one referring to semi-historical beginnings of 
society and mythical personages, the other to actual genea- 
logical facts amongst Vedic (=Epic-Puranic) ruling -and 
priestly families, some of them comparatively late. The 
former class may admit of mythological interpretations, 5 
though that does not explain why such parental incests should 
have been favourite similes and been at all ascribed to persons, 
historical or legendary. The second class is evidently historical, 
and certainly was not the product of a primitive and barbarous 
community : the Vedic civilization proper had already reached 

1 Mbb. 553 (Yalvah. ) : IT, 70-72, 2267 ff. 

2 Gf. the many indications in Mbh. IV, 35-37; e.g., Krsna's hint to 

Uttara : " Arjuna will doubtless obey your sister of graceful 
hips 1 '; voluptuous description of Uttara's approaching Arjuna in 
fclw dancing-liallj and making her request to her * sakha f , dis- 
playing ' pranaya ' and coquetry, * like a she-elephant seeking 
her mate ', vowing suicide if he did not keep her request, etc. 

3 Vide n. 2 S above. She is among the circle of court-ladies attending 

on the^ gay, spoilt and musical prince Uttara. Her developed 
youth is described j and she bears a son about six months after 
her marriage ^ a few days later. Playing with dolls, is still a 
common pastime with grown-up girls in many social circles in 
India, often continuing far into their married life; so also the 
post-Mauryan ' Kagaraka * (in Vats. Ka. Sut.) captivates his lady- 
love as much by presents of dolls as by taking her to clubs 
dances and theatricals. This in itself, therefore, is no reason 
for concluding child-marriage in Uttara's case, as the Cambridge 
Hisfe. of Ind. does. 
I CL the parallels of Pratipa, Bhagirathi and Santanu ; and Jyamagha, 

hi* captive maiden, and Vidarbha ; vide infra. 

Such explanation is not seriously attempted by the Puranas; the 
Branmanas do it, and that because the Bgvedic reference itself 
is a mythological version of a Puranic tradition. (It would seem 
as if the Alias generally mythologised the traditions of the pre- 
existing peoples, viz., ^he Manvas and Prthuites.) 


its highest point/ when these cases are indicated, and the 
last phases of the Epic age were being worked out. Thus 
such connexions between parent and offspring, or persons in 
equivalent position, cannot have been due to the needs of a 
strongly patriarchal, primitive and -conquering community; 
they were rather the extreme result of two well-known forces 
that have worked amongst various early 2 but civilized peoples 
the tendency in long-established hereditary priesthoods and 
ruling families to continued in-breeding, " and to unlimited 
license. As a matter of fact the close of the Yedic age, which 
is the same as the Epic age, shows evident signs of increasing 
social degeneracy in many other respects, which clearlv con- 
tinned till the tune of the Upardsads and the development of 
the great Reformation in the Praci. 5 This general outline 
will emerge again and again in vie,w as we proceed to examine 
the evidence in regard to other social details. 

*u T lJegard to the natllre of the aster-marriages also, 

the Yedic ; evidence by itself suggests no very adequate explana- 
tion of the references to them; and here/ again, the ' tradi- 
+ OI T- ,? vldeBce ls somewhat helpful. An examination 5 of 
the Vedic uses and imports of words designating brother and 
sister, and of their comparative position in the Vedic (Brah- 
mapical) family, as indicated by incidental references, yields 
rather uncertain Results : these uses and indications only make 
it passible that sister-marriage may have developed in a com- 
munity ^ and age, which was either strongly patriarchal and 
emphasised the brother as master and supporter, or which bein s 
originally matriarchal, still retained traces of. the importance of 
the sister in the family and descent through the mother. 5 

1 Between Mandhat* and Sudas roughly, about 20 steps before the 

-fcSnarata period, - 

2 Thus Artaxerxes Mnemon (early 4th cent. B.C.) married his daughter 

Atossa: Sykes : Hist. Pers. I. 246, Medieval and 
history id left out of account here. (Ihis tendency Is found alo 
in small communities with a hereditary occupation : .thus more 
or less confianguinouB marriages are not infrequent amongst certain 
modern trading castes in Bengal.). 

3 The check, however, sterns to have been only temporary; for m s i- 

Mauryan morality (cf. Vats. Ka. Sat., re dynastic and priestly 
customs) is <pte as bad as pre-BuddMstic ,- indeed, the evidence 
of the Arthasastra would seem to show that within non-Buddhistic 
spheres of influence there was little change in tone even in the 
early Mauiya period ; probably the only pnritanistio age for the 
whole oonntay was that of Asoka, and that of the growth of 
early Buddhism from before his time, in limited areas. 

4 Except that the sister (and specially the twin-sister) was supposed in 

early times to be the brother's wife by birth-right (Rv. X, 10, and 
Ar. XIV, 2, 53), and that such connexions were sanctioned in 
case of necessity for a son and heir (Ev. X, 10, and the * gatha * 
in the gnnahiepa legend) : both of which indi<cationg point to a 
patriarchal origin. 

5 Vide ante. 

6 Vide ante 

( 144 ) 

The independent value of such linguistic evidence in history 
is rather doubtful. The Puranic evidence makas the ground 
somewhat clearer : while there are two probable oases of one 
sister marrying two brothers, 1 there are definite as well as 
probable cases of a brother marrying two, three, or more 
sisters 2 ; and in other instances the .sister is ^only one of 
several wives. 3 Hence the noticed sister-marriages in the 
Parana- Vedic period were rather more patriarchal in ^ features 
than matriarchal, being more definitely connected with poly- 
gamy than with polyandry. On the other hand, some of the 
early instances show that the chief part in such unions was 
played by the sister 4 ; and the two apparently polyandrous 
cases were also comparatively early. 5 Hence the matriarchal 
type of sister-marriage was the earlier one. It looks like 
having an ethnic significance. But the references in view 
belong to all the three broad groups of the Prthuites, Manvas 
and Ailas, though chiefly to the non-Ailas. 6 Indeed, the 
selection of the Manva. case of Yama and his sister as typical 
in Ev. X, 10, would indicate that so late as the date of that 
' vakovakyia.,' the sister usually took the initiative in such 
connexions ; though on the other hand the earlier Ev. VI, 55 
would suggest that it was the brother who took it ; and one 
of Yami's motives is to have the full extent of a brother's 
rightful * protection J and bear a worthy gnandson for their 
father (i.e., a pure-blooded one), a patriarchal trait. 7 

Thus the Vedic sister-marriage must have originated in 
two distinct pre-Mstoric types of civilizations, which blended 
their features in one, probably to be indentified with the Aila 
and the pre-(and non-)Aila. 

1 At the same <fcime, or (apparently) In succession : with Haimavati- 

Dr^advatf and ^Narmada, respectively, both in the Aiksvaka line ; 
cf. similar indication in Sita's case (vide ante), also in th same 

2 Bharata (Angirasa) : 3; Bhajamana and Satrajit (Yad&vais) : 2 and 

3 E.g., with Das'aratha (Aikvaka), or Brupada (Pancala), or with 


4 E.g., Snnitha, Yami', Acchoda, Narmada : vide ante. 

5 About 70-74 steps before the Bharata battle. 

6 The instance of Acchoda alone being an Aila one; Sunitha is 

iPrthnite; Yam! and Narmada,, Manva, to which may be .added 
HaimavatrDrsadvati, for Prasenajit is known as her son (vide 

7 This is also the motive in the earlier Anga-Sunitha case. 


The Yedic evidence 1 suggests that polyandry was not 
altogether 4 un-Vedic ' ; it was apparently known, though 
particular instances are not named, which' silence has at best 
only a negative value, for full details of these matters cannot 
be expected from the nature of the Yedic literature. The 
Epic tradition definitely assigns polyandry to the close of the 
Yedic age ; and very much earlier, even pr-e- Yedic 2 instances 
are known to Puranlc and Epic tradition. The number of 
illustrations of peculiar customs is naturally not large, 
specially as later editors were busy; in removing striking traces 
of primitive characteristics that had become offensive, A re- 
markable case .of such removal is that of the polyandric 
marriage of a brahman lady, Gautami : the Epic affirms that 
in the time of the Pandavas one authoritative precedent of 
polyandry was that of Gautami, who married seven *rsis, f 
and that the case is recorded in the Puranas 3 ; but the Purana 
texts, in their royal or priestly genealogies, have no such 
mention now : obviously the instance has been removed In 
brahmanic interest. 4 In the cases of sister-marriages and 
incestuous connexions, it has already been shown how texts 
have been emended, muddled, misinterpreted and mythlfieS, 
wherever prominent examples of these were noticed;" in the 
case of polyandry, as well, the explicit instances that have 
escaped weeding out and emendation are few, but it is still 
possible to see that many more were known at one time, 

Before proceeding to examine these probabilities, and the 

the famous epic instance, which was too well known and 
late to be successfully buried, 5 the two explicit references 

may be noted here. The ten grandsons (or great-grandsons) 
of the famous Prthu-Yainya married a common wife Marlsa, 6 

1 Tide ante. 

2 I.e.. Prthu-ite, being several steps above the earliest group of Manva 
rsis in the Rgveda; Prthu-Vainya is, however., also Included within 

the Vedic anthology ; and the case referred to is assigned to three 
steps below Prthu. 

3 Mhh. 237 (Vaivahika. ) : I, 196, 7265. 

4 While the non-brahmanic case of Marlsa m. 10 Havirdhanas (or 

Pracetasas) referred to in the Epic in the same connection (Mbh. 
I t 196, 7266) is found in all Pur-anas. 

5 Even here, cf. the explanation in the Markandeya, that it was really 

a c monandry *. since the five Pandavas were parts of the same 

5 Visnu : I, 15, states that Marisa in a former birth became a childless 
"young widow, and obtained a divine boon for several husbands 
a|> the same time to ensure non-widowhood and progeny. 

( 146 ) 

a daughter of Soma. 1 The Puranic account further specifies 
that this happened, because in the Caksusa- 1 Manu's ' 
period 2 the population or dynastic birth-rate declined, and 
those ten princes, the Havirdhanas (also known by the 
common appellation Pracetasas) , 3 were admonished by Soma 
to procreate, who gave them his daughter Marisa as their 
common wife; 'they' had by her Baksa, the ' prajapati,* 
who was very prolific, 4 and other children also, 5 but no 
' fathers * are specified in any case : Daksa-Pracetasa in fact 
is often said to have had tea * fathers.' 6 The other explicit 
mention is about the brahman lady, Jatila-GTautami and her 
seven ' rsi ' husbands, 7 Her example must have been well 
known and appreciated at one time, for in Mbh. (besides 
Pandu's reference) the wives of citizens admire Draupadi in 
the company of her five husbands and compare her to 
Gautami with her * rsi ' husbands. 8 The chronological 
position of this case is not so evident, but the outside limits 
can be fixed: she cannot be placed before the Gautamas are 
first mentioned 9 in Bharata's or Marutta's time, or later than 
the Pandavas, to whom she is a precedent ; and there are 
some indications in favour of the earlier limit. 10 

Taking the less definite cases, inferable or probable, in 
chronological order, we come first to the already noticed 
combined polyandry and brother-marriage of Haimavati- 
Drsadvati, in the 18th step 11 from Mann and in the Aiksvaka 


1 As her son's daughters were also married to a ' Soma ' (in all 

accounts of the pre-Ailas), it would seem that ' Soma, ' was a 
clan name even before it was used to designate the Aila dynasties 
derived from ' Soma ' ; cf . the carious question on this point in 
Visnu : I, 15, 80-81, 

2 I.6., the interval between the 6fch and 15th steps in the Prthuiie 

dynasty, and between the 3rd and 12th steps before the 
Vaivasvata-* Manu *. 

3 Mateya : 4 (Svayambhuva genealogy). 

4 Harivamsa : 2, 88-106 ; Mbh. 137 (Sambhava. ) I, 75, 3130. 

5 Matsya : ibid. ; viz., Nandi, Candravati, etc. 

6 Mbh. I, 53; 3130; 75; etc.; cf. Hariv. V, 66 fl, and Mbh. S 665 

(Moksa.o) : XII, 208, 7573. 

7 Yide n. 3, p. 145. 

a Maharsln ive- Gautami ' : Mbh. 635 (Bajadh. ) : XIT, 38, 1397. 

9 Utathya-Xngirasa beingr regarded in the * rsi ' genealogies as the 
first Gautama a or Dlrghatamas, his son, "acfcording to other ver- 
sions (cf. Pargiter : AIHT. pp. 219-220). 

10 Vide infra. 

11 About five steps above this, in the time of Kuvalas'va-Aiksvaka, 

the contemporary Paurava Sudhanvan-Dhundhu (made into, an 
Asnra adversary) is said to have been son of two brothers, Manhu 
and Kaitabha (Mbh. $ 475 Dhundlm ) : III, 202 3 13532; G4, 
13587); this looks like legend j but it seems likely' that 
1 Manasyvabhayayoi putram * of some dynastic sloka has been 
made into * Madhukaitabhayoh putrarn ', k> remove odium from 
the Faurava dynasty : Manasvn and Abhaya were the immediate 
predecessors of Dhundhu, If this view may be taken we have 

( 147 ) 

line; and only four steps lower, lo the almost parallel case of 
Nannada (m. Pmiikutsa and Ambarisa), where the some- 
what meaningless ' sambhuta ' of the texts (Instead of being 
a proper name) may be a- relic and a corruption of ( sam- 
bhfiya,* san-mata<t, s or * sammata/ 1 referring to Xannada's 
being wife of Purukutsa and brothers in common. 

About 19 steps later on, there is the much clearer case 

of Mamata, amongst the first G-autamas (a section of the 

Angirasas) ; and Jatila-Gautamf s polyandric marriage (cited 

by the Pandavas) must have been due to a tradition of such 

marriages in this family. Mamata 2 is said to have been 

Utathya's wife, but his brother Vrhaspati had free access to 

and equal conjugal rights over her in Utathya's life-time ; 

the only objection Mamata once raises to t&eir exercise is 

her pregnancy at that time; she asks him to wait, but does 

not refer to any impropriety or unlawfulness of conduct; 

evidently she was in the status of a wife to both brothers. 3 

So also Vrhaspati and Mamata' s son Bharadvaja, is said to 

have got that name from the circumstance of his being 

4 born of two fathers/ who both charged the mother 

Mamata with his maintenance ; the derivation may be an 

ingenious after-thought, but the fact referred to is "original. 

This Bha-radvaja is also called ( dvamusyayana/ which is 

usually explained as referring to his adoption by Bharata, so 

that being the son of a priest by birth, and of a- king by 

adoption, he would be the son of ' two fathers/ But the 

details of that famous tradition 4 of Bharata's adoption show 

that It was not Bharadvaja himself who was adopted, but 

his son or descendant Yitatha for Yidathin), who seems to 

a parallel instance in the Paarava line as well, Tiie probability 
Increases when we find a Sainyati section among the KSsyapa 
brahmans (Matsya : 199), who counted a number of ( dvamusya- 
yana * or blandric families amongst them (vide infra.) ; and 
Samyati was a near successor of Dhundhu. (It -was, cf course^ 
common for princes to found rsi or brahman families affiliated 
to different gotras). 

1 Vide collated test, ante. ; cf. the epithet ' sammata bharya, * of the 

Pan^avas, given to Draupadi, and ( satam matat ' in the case 
of Haimavati. 

2 For these details re Gautama family (connected with Bharata and 

Vali In tradition), cf . Matsya : 49, 11-34 ; Visnu : IV, 19, 5-8 ; 
Matsya : 48, 32-57: Vayu : 99; Brahmanda : III, 74; Mbh. 170 
(Dlrgh, ) : I, 104. ' 

3 It is noteworthy that Tara, the_ wife of a much earlier Vrhaspati. 

also stated to have been^an Angirasa, was desired by his brother 
Dharma, who however did not get her, being obstructed by her 
paramour Soma (Varaha : XS2II). Though rather semi-legendary. 
the tradition certainly is of value as showing trace of polyandry 
among Angirasas (to whom other primitive forms of connexions 
are also ascribed). 

4 I?or these and other connected details dealt with here, cf. Brahma : 

IS, 58-60; Matsya: 49, 11-34; Visnu ; TV. 19, 4-8; 
99 j Hariv. 32, 1726-'ol 

( 148 

have been really a ksetraja son of Bharata through Sunanda 1 ; 
probably it was no case of adoption at all: f samkramana ', 
transmission, grafting or infusion, may equally refer to a 
1 niyoga ' 2 . So Bharadvaja was a ( dvamusyayana ' in some 
other way, evidently because he was ' born of two fathers/ 
Utathya and Vrhaspati, whose joint wife Mamata was, in 
the same way as Daksa was, l son of ten fathers.' Thus we 
find, besides descendants of Bharadvaja, three other Angirasa 
and eight (or twelve) Kas'yapa families 3 designated 'dvamusya- 
yanas ' ; all of their forefathers cannot have been similarly adop- 
ted by childless kings, and they have no evident connexions with 
any dynasty; but these brahman clans may well have had 
some sort of a biandric custom 4 originally, 'it is noteworthy 
that in the nest generation also , the same features are repeat- 
ed to some extent. 5 Thus Dirghatamas freely approaches 
his younger brother's wife 6 ; and like Mamata, Dirghatamas's 
wife Pradyesi maintains her children, even the husband; 
and his ruling on her (and on all women thenceforwards, it 
is said) restricting her to one husband, shows that she too 
(probably like other Angirasa women) followed Mamata, 7 as 
Dirghatamas followed Yrhaspati. 

1 Mbh. states that as a result of Bharadvaja's good offices, Sunanda, 

the queen of Bharata, bore Bhumanyu, after the nine sons had 
perished : Mbh. 151 (Puruvam .) : I, 94, 3710 ff. 
'Yr^ 1 ' 0111 a cansideratio tt of all the traditions about Dusyanta, 
Marutta, Bharata and the Gotama-Angirasas, it is clear that the 
'samkramana ' of Bharadvaja was due to the influence of 
Marntta's family; Marutta's daughter Samyata was given to his 
Angirasa priest Samvarta, brother of Utathya ; it is possible that 
the interest of the Marutta-ites in this adoption was due to 
Bharadvaja's being born of this princess, who may well have 
been the common wife of all three brothers, and the same as 

2 Vide infra, sec. re ' niyoga '. 

3 Huta,_ Samiga and "Saisira.- Angh-asas (Matsya : 196, 52)- for 

Kasyapas, Matsya : 199, 11-12 (Saisira being common . 

4 As amongst the Manvas and other non-Aila peoples ; vide infra 

(Kasyapas are probably = Manvas; the name Kasyapa itself may be 
of Dravidian origin; so also the name Angirasa). 

5 Of. n. 2, p. 147. ' 

6 Though he is cast out apparently for thus transgressing the limits 
of an ' elder ; brother, the main objection against him was his 
passing the limits of decency in other ways, and it was more 
bis wife than his brother ^ho banished him, At most we have 
nere probably an intermediate stage in the development of 
polyandry, the wife of the ' elder > brother only being common 
to the younger brothers, but not vice versa (as also in the case 
of * niyoga ' and widow-remarriage, where the rights of the elder 
brother were restricted subsequently). Cf. one o? the objections 
raised by Dhrstadyumna against Yudhisthira's marrying Braupadi, 
who, having been won by Arjuna, was virtually an younger 
brother's wife. 

l t Mamata and Pradvesi's economic position in*the Gautama familv 
is evidently a trace of a passing; matriarchal custom ; cf . the 
metronymie Mamateya j cf. also the mother as ' bhartri in Ved, 
lit. (vide ante). 

( 149 ) 

Polyandric traits crop up again In tradition about 
20 steps further down, and all In the same connexion. It is 
noteworthy that these refer mainly to the Deccan peoples 
connected with the Manvas. The Eamavanic ! tradition 

V * 

(common to both the epics and the Puranas) affirms this 
feature of Kiskindha, 1 where Vali and Sugriva are born o! the 
same mother Viraja, wife of Eksa, by two co-existent 

paramours 2 (?), and they, in turn, practically had either the 
wife Tara, or the wives Tara and Eiima, in common, though 
they quarrelled about it and excluded one another alternately. 3 
Further south the relationship between ilandodaii and 
4 Havana ' and Yibhisana 4 Indicates a similar polyandric trait, 
over and above ' devr '-marriage. It Is quite possible that in 
1 Surpanakha- ' attending on her brothers 5 during their early 
austerities, it Is a case of combined polyandry and sister- 
marriage : for the onlv other ' traditional ' Instances where 


austerities are assisted by an attendant woman are those of 
Agastya and Lopamudra 6 and (the legendary) Siva and Uma, 7 
in both of which the woman is the wife. It Is also signi- 
ficant that it is only the ' Baksasa ' chiefs of the S.E., who 
hunt or roam about accompanied by a sister, who often acts 
independently, and excites the resentment of and endangers 

1 Probably it Is needless to say now that the Vanaras and Raksasas 

represent real races, perhaps in some way connected "with later 
Dravidians and Kolarians. with occasional Aryan admixtures. 

2 E.g., Brahmanda- : III, 7, 212-16; etc.; cf. Mbfa. Ill, 147, 11193 f. ; 

Earn. VII, 42. " 

3 E.g., Padma : IV, 112 (Pur. Ham. ) : 146-165 (Brahmanda III, 

7, 218-21 names Tara and Ruma, but omits the fraternal 
strifes) ; cf . Ram. IV (Kisk.) : Taravakyam, or sees. 5 to 55 
generally, and sec. 46. 

4 Cf. * Tara Mandodari tatha * in the traditional couplet about famous 

polyandrous women of history. With Mandodari it was apparently 
also a case of brother-sister or cousin marriage; for she describes 
herself (Ram. VI. 113) as a daughter's daughter of Somali, who 
was also the maternal grandfather of Ravana ; Mandodari's mother, 
the light-skirt Hema (who had a disastrous amour with M.'s 
father) was thus either the same as Rav.'s rather forward 
mother Nikasa (Kaikasi), or her sister. 

5 Twin as well as step : Mbh. Ill, 275. For the possibility, cf. Bam. 

Ill, 21, where gurpanakha calls Khara her ' natha/ and he too 
speaks of himself as her * natha.' It is to be noted that Siirp. 
concealed her love for Rama and Laksmana from her c natha * and 
Ravana, and invented reasons for her plight. 3urp. is said to have 
first been married to the Kalakeya Vidyujjihva, but Rav. killed 
him in battle, and then made her over to his brother or cousin 
Khara, with whom she continued to live, obeyed by him (Ram. 
VII. 29). To Rama she said : " Passing over (ati-kranta) my 
brothers Ravana, Kumbhakarna and Vihhisana, and the two 
brothers Khara and Busa-na, I am approaching thee as husband, 
falling in love with thee at first sight, so be thon my husband 
for Ions " (Ram. in, 17). 

6 Fadma : vf 2^ 40-1; cf. Mbh. Ill, 97, 8579-80. 

7 In the later sees, of Pur. and in the Kavyas : Matsya : 154-158 

(the germ of 'Kumara*); Varaha : XXI-XXIIj etc. 

( 150 ) 

her brothers, by her sudden .and misplaced loves. 1 The 
fraternal polyandry of Sunda and Upasunda also seems to 
belong to the" generation before Rama, and to the N.E. part 
of the Deccan, 2 The Mbh. illustrates 3 the danger of polyandry 
by the famous story 4 of these two chiefs of the Vindhyan 
uplands quarrelling over the same woman; and even before 
Tilottama's appearance, the two brothers seem to have had 
other women in common, but without any resultant troubles. 
Considering all this ' Kajnayanic ' evidence, and the already 
noticed polyandric (and biandilc) traces amongst the Aiksvakas 
(Maims) and connected groups like the Angirasas, etc., 5 it 
seems not unlikely 6 that, in the original tradition, Sita was 

1 Cf. the stories of c Surpanakha and her brothers ' Bavaria,' etc., 

and Khara, etc., and * HioTnnba ' and her brothers (Hi^imba 
And Vaka, etc.), in different sources and periods. I agree with 
Pargiter in thinking that ' Surpanakha ' and 'Hi^imba ' are 
Sanskritised forms of the original Dravidian and sensible epithets 
of * Surupnagai ' (ruling or crown princess) and ' loUmba ' (proud 
woman or empress); so- also ' Bavana ' = ' Iraivan ' (lord, king), 
and * Hanumant '=* Anmandi ' (male monkey = * Vrsa-Kapi ), 
a. patron deity of the Dravidian Vanaras, or perhaps even of 
Kosalas; elsewhere I have suggested that brahman gotra names 
like Angirasa or Kasyapa may be Sanskritised forms of Dravidian 
clan names (meaning * magician * and ' mat-seated father/ 

2 In Earn. Marica is son of Tadaka (a non- Aryan chief tainess of 

Malaya and Karusa) by Sunda (a descendant ot Dhundhu), who 
shortly came by his end; and though not a pure Baks.asa by 
birth he cam& to be regarded as such. In the Pur. *, of the 
two brothers Sunda and Upa(Ni)sunda (vaguely derived from 
Diti's race), Sundays son by Tadaka was Marica, while one 
reading seems to have implied that he was c born of Tadaka from 
Sunda and Upa(Ni)sundaf (with which may be compared 
DhundhUj son of ' Madhu ' and ' Kaitabha,' ante}. The 
geographical setting of SundOpasunda's story in Mbh. agrees with 
that of Marica and his parents in Ram. , being the same 
Vdndhyan forests and tablelands bordering on the Gangetic 
valley. According to Mbh., Sunda and Upasunda raided the whole 
country from their Vindhyan home (cf. similar devastation 
attributed to Sunda's family in Barn. ) and reached Kuruksetra, 
which is quite proriable, as at this step in the dynastic lists the 
Kuru kingdom was in abeyance owing to Pancala raids (cf . the 
circumstances of the Eaksasa occupation of Varanasi) ; hence the 
story of Sunda and Up&sunda's destruction through a biandric 
practice must have been well known in the Kuru country, and 
the allusion to it in Pandava court is "therefore genuine. 

* E.g. Brahmanda : III, 5. 34 ff ; Vayu : 67, 72-3, 

t Thus * Nisunda J is &n alternative for * putrastu, 9 - Vayu : o^. 
cit. ; proh. in Brahmanda op. cit. the true reading might Joe 
' Marico SaundOpasundas Tadlakayam ajayata J instead of '. . . 
Sunda-putrastu. . . '. 

3 Vide latter part of last note. 

4 Mbh. 246 (Eajyala. : Sundop. ) : I, 209-212. 

5 Vide ante. 

5 These probabilities need not upset admirers of the epics, for the 
actual events of the Barn. occurred at least 1,300 years before 
their Kavya idealization (which process indeed has continued 
through the middle ages to the present day), and the ideals cf 
subsequent ages of course do not suffer. 

151 y- 

the comuion wife of Eama and Laksmaiia, just as it is clear 
that she was originally the sister- wife of Eama\; indeed, the 
episode of Laksmana refusing to go to assist- Kama while MB 
cries of distress are heard, and Sita charging him (and 
Bharataj with a design of appropriating herself after getting 
rid of Kama, 2 seems to point to this original relationship, 
which would then be paralleled 3 by the case of Tara and her 
husbands at strife, amongst a people friendly and probably 
kindred to the Manvas. 4 

For about 25 steps after this, tradition supplies no trace 
of polyandry (or biandry). Then, again, indications become 
evident during the several generations before the Bharata 
battle. It would almost seem as if these apparent recru- 
descences are due only to the variation of the tradition in 
fulness of detail, and are not real reappearances. 5 

In connection with the Pandava proposal of poiyandrj 5 , 
indeed, Drupa-da is said to have been shocked at its novelty; 
but Dhrstadyumna gives the whole show away by arguing 
that YndliistMra as elder brother of Arjiina could not marry the 
girl won by the latter, thus showing that a- rebLricted polyandry 
was known to the Panoala court 7 ; and KrsnaJDvalpayana 
further spoils the case, by explaining how the practice was 
established and is to be recognized, and one of his two 

1 Vide ante. 

2 Earn. Ill, 45 and 49. This original relationship seems to be con- 

firmed further by Rama's suggestion that. Sita might live as wife 
wiih Laksmana, Bharata and atrughna (Bam. VI, 117), and by 
Viradha's surmise that Sita was the common wife of Kama and 
Laksmana (Bam. Ill, 2). Eama had proposed Sita's transference 
to Bharata even before her abduction, on the eve of his exile 
as a convenient arrangement during his absence (Ram. II, 30, 
8-9 7 with 26, latter part). 

3 Cf. also the case of Nala, Puskara and Damayanti In Upper Deccan, 

a few steps above ; also that of Mandodari (virtually a Manva 
case, for Ravana's line was traced from that of Vaisali) ; cf . 
Vali and Ravana vowing to have wives in common (like Sugriva) 
as a token of friendship : Ram. VII, 39. 

4 The Manva families of Ayodhya, Vaisali, etc., and the brahman 

families of Angirasas, Kasyapas, Vasis.thas, etc., were apparently 
originally Dravidian (at any rate extra- Aryan traits are found ' 
largely amongst them, though the Aila^ are not altogether free 
from them). The comparatively later and wrong legend of M-itra- 
Varuna and Urvasi seems to indicate an original custom of biandry 
amongst the Vasisthas, as amongst Angirasas and Kasyapas, like 
whom they also might be called * dvamusyayanas-' 

5 So also with regard to tHe reappearance of other forms like sister- 

marriage, etic. 

6 Mbh 237 (Vaivah. ) : I, 195, 7226 ff. ; 7255-7263. VySsa's explns. : 

238-'9 (Paficendrop, ) : I, 197, 7316 if- 

7 So also Vidura is said to have advised householders of Indraprastha 

and Arjrraa's successor there to desist from polyandrous marriages : 
Cowell : *Jat. VI, 139, etc- ; vide infra, pp. 161162 for Jataka 
yersion of the Pan^ava polyandry (on Krsna's ovsn initiative). 

( 152 ) 

explanations, shorn of fable, plainly indicates that even in the 
nest previous generation polyandry could occur in a good 
* rsi ' family 1 (while the other explanation seems to refer to 
a dynastic case) . 2 Above all, 3 even before the ' svayamvara * 
of Dranpadl, Ki'siia-Dvaipayana takes the polyandry for grant- 
ed as an ordinary thing supported by \rsi ' precedent, and 
advises the Pandavas accordingly, twice 4 ; and KuntI is 
remarkably insistent in her demands all along, all that 
explanation by her ( dread of untruthfulness ' being evidently 

This last point, and the fact that, whether by way of 
niyoga ' or by way of polyandry, 6 Kunti had herself known 

1 The polyandric tendency, often amounting to unrestricted license. 

lingered on amongst the "brahman families, even after the 
Bharata battle, specially among the Angirasas, Kasyapas and 
Atreyas : vide infra. 

2 Vide infra. 

3 For other indications of frequency of polyandry, vide infra. The 

very fact that Draupadi was able to conceal her identity by 
professing to be the common wife of five ' gandharvas ' (what- 
ever may be the real meaning of that term, ' Kinnaras '= Upper 
Sutlej Mllmen, or simply professional musicians or Kusllavas, 
whose wives, according to Vats. Ka. Sut., are not confined to one 
husband);, shows that polyandry was fairly well known in the 
Matsya country as wH. (Probably even ' gandharva * is an after- 
thought, and the Pandavas in their incognito exile simply 
passed themselves off as another humbler polyandrous fami'y; 
vide n. 7, p. 151). Drau. is taken to be a gandharvi \v. of the 
sons of a gandh. king (Mbh. IV, 9, 257). She professes to be 
w. of five gandh. (IV, 9, 273 ff.; 14, 426; 16, 493 ; 21, 664; 
22, 787), So also the Pand. are mistaken for gandh. (Bhi. : 
IV, 8, 235; 22, 792; 23, 819; 71, 2293; Nak. : 12, 323; Arj. : 
45 3 1406). Note that it was a * gandharva ' who advised 'the 
Pand to contract a polyandrous marriage with a Kasyapa priest's 
help, and that gandharvas were Kinnaras (Mbh. II 5 10, 396 : etc. ; 
vide Sor. Index, s.v. gandh. and Kinn.), among whom Pandu 
lived and allowed Kunti's and Madri's five connexions, and whose 
modern representatives the Kanwaris are still polyandrous. 

4 Mbh. 220 (Caitraratha ) : I, 168 ; 169, Dhaumya was chosen (at 

the instance of a Gandharva chief) the Pandava family priest 
in view of their intended polyandrous marriage, and he per- 
formed their nuptial rites according to a form whereby the 
common bride was deemed to have regained virginity after each 
individual marriage and its consummation; he also performed 
the usual ceremonies for the children of this marriage ; his 
kinsmen were also the royal chaplains of the Pancalas (Mbh. 
I,_ 183; 198, 7338; 221, 8047; etc.). Thus the Gautama 
(Angirasa), Vasistha and Kasyapa brahmans were all familiar 
with polyandric marriages, as much as the princes and people 
of Indraprastha, Matsya and Pancala. Cf. the regret of 
Draupadi that she and her husbands were not born as brahmans, 
for amongst Ksatriyas she had been called a cow by Duryodhana 
for her polyandry; Mbh. 340 (Ariunabhig. ) ; III, 37. 

a Mbh. I, 196. 

6 Vide infra, 

( 183 ) 

several husbands,' make it likely that polyandry 1 ^was then 
also known amongst the Yadava races. 2 The uniform state- 
ment in the Puranas that 'Ahnki' (three steps above Exsna) 
'was given in marriage to the Avantis (or Avanti princes)', 5 
also Yadava s, may refer to this lingering practice ; the form 
of the .statement is too unusual in the genealogies to admit 

tf~^j t? 

of any other meaning. In the Epic and Puraiias the Avantis 
have two co-klxigs, 4 in the third step after Ahuki; Yidarbha, 
another Yadava state, was in the same period ruled by joint 
kings, apparently representing two sections of the same 
dynasty, the Kratha and the Kaisika 5 ; Magadha in the same 
period had a succession of dual kings 6 ; so also in Easkindha, 
where ' Mainda ' and * Dvivida * ruled the kingdom of Tail 
and Sugiiva. 7 If these instances of ' diarchy/ in the same 
age and in a continuous belt of country (the Deccan and its 
borderlands) , were not purely accidental, they may have easily 
led to a dynastic custom of having a common k mahisi J by 
way of ' biandry.' 8 Apart from this possibility there is surer 
indication that tradition knew of kings of difterent dynasties 

1 Along with other primitive forms. 

2 it was apparently also known amongst the people of Alahismatl and 

its .Pauiava printees; the custom of sexual liberty of Mah. 

wives who were not confined to one husband was noticed by 
fcsahadeva Pandava when he conquered that kingdom $ it- was 
said to have been sanctioned or established by brahman ordinance 
(Mbh. II, 31, 1124-40); ci. the tradition about Uddalaka; also 
the w. of an k Atri * (pro-Yadava and Central Indian) leaving one 
husband and haying issue by another agent : (Mbh. XIIJL 14. 
684, f.). 

3 s Avan&ibhyah ' ; one text emends to l Avantisu * ; probably ' Avanti- 

bhyam ' would be a good reading (vide nest note) ; Brahma : 
15, 48; 54; Hariv. 38, 2017; 2023; Matsya : 44, 66-70; 
Brahmanda : III, 71, 121 ; 128 ; the Vayu text is corrupt, but 

obviously its source was in the same form as other tests. 

4 ' VindAnuvindau ' : Vayu : 96, 145 ff. Brahman^a : HI, 71, 150 fi. 

(confusing with the two Kekayas of same name mentioned in 
the Epic) ; Matsya : 46, 3-10 ; Visnu : IV, 14, 10-11 ; same 
in Hariv. ; Padma : V, 13^ 56. 

5 The brothers ' Kratha * and Kaisika * were the joint rulers of 

Kundina Gity : Hariv. 108, 5980-81 ; Bhlsmaka being the 
' Kai'sika ' in Krsna's time (often in Hariv. 105 to 108) and 
Akr (hvr)ti being the * Kratha J (ibid., sp.^in the lists of kings 
opposed to Ivrsna in connection with Rukmini's abduction), 

* In Mbh. : cf . Sor. Index (p. 355) for their names, chiefly, JarasandJaa 
and Jalasandha; Jayatsena and Sahadevaj Danda and Dan^a- 
dhara, etc. 

7 Mbh.; in the account of Sahadeva'a southern campaign; and Hariv. , 
in that of Krsna's exploits. (These two names were probably 
dynastic ones, *as they also occur in connection with Rama's stories 
and in Puranic ' Vanara * genealogies). 

9 As earlier m the case of Vali and Surgriva; cf. MandodarL 

( 154 ) 

sharing tjie favours of a princess 1 by agreement : the apparent- 
ly wild tales of ' Yayati 's daughter ' 2 and the ' five Indras ' 3 
prove this. 

This former story takes us back to a period 4 before the 
4 Gautaana ' cases of polyandry, quite an early stage ; and is 
told -of persons who are otherwise 5 famous in tradition ; hence 
the amount of fable and brahmanical edification that has 
entered into the account 6 is only what might be expected, 
specially as the behaviour of those personages was far from 
creditable. 7 There are some obvious historical mistakes in the 
story due to subsequent brahmanical handling, 8 but their 
sources can be discovered 9 ; some of the persons named as 
contemporary are clearly so, 10 while about others there Is no 
direct traditional evidence to the contrary 11 ; and the story as 
a "whole is referred to in other connexions and finds support 
from incidental Vedic, Puranic and Epic allusions. 12 

I A probable case of such sharing (though not peaceful, apparently) 
is indicated among the Yadava-Pauravas of the S.W., a 
generation before the Bharata battle : the King of Karusa 
(either Yrddhasarman or Dantavakra), Sisupala of Cedi, and 
Vasudeva of Dvaravati (and Mathura), are all staled to have 
had Bhadra-Vaisali (which name lean have belonged to only one 
person) for their wife. (Sisupala, however, obtained her by 
impersonation or force,* but regarding Vasudeva and Karusa 
there are no special statements. This Bhadra is also stated to 
have been Sisupala's maternal uncle's wife, whom he enjoyed 
under the disguise of the Karusa king, who was ihis mother's 
sister's husband. So Vasudeva and his brother-in-law apparently 
had equal access to Bhadra- Vaisali. She however subsequently 
ascended the funeral pyre of Vasudeva). Vide Mbh. 291 
(Sisupala. ) : H, 45, 1570 3 ; 793 (Ma-usala. ) : XVI, 7, 194 ; 
cf. Brahmanda : III, 71, 173-4, and corresponding passages in 
other Fur. 

S Mbh. 565 (Qalava. ) : V, 114-120. 

3 Mbh. 238 (Vaivahika. ) : I, 197. 

4 Befoie the 40th step and after the 20th step from Mami (which 

latter is the date of the beginning of Haihaya, raids). 

5 Specially in connection with the Haihaya invasions. 

6 So also in other stories told about &ivi ? Pratardana, Vis'vamitra ; 

or about Sagara, the Bhrgus, etc. 

7 So also the Pan<Java polyandry is cloaked with ill-fitting puerile 


8 E.g. in making Galava the central figure of the story , or Yayati a 

contemporary of the four kings. 

9 Thus Yayati may easily have been substituted for Ahamyati or 

Samyati, who were contemporary Yayatya kings at Pratisthana. 

10 E.g. Visvamitra and Samyati (through Krtavirya) ; Usinara and 


11 E.g. re Usinara, Divodasa, Haryas'va and Visvamitra. 

12 E.g. Mbk III, 197, 13301-2; I, 88-93; V, 119-122. Matsya : 35, 

5; 37-42. Ev. X, 179; III, 31, 1-3 (by a Visvamitra or 
Kusika rsi) seems to refer to and justify Haryasva's begetting 
a famous son for the benefit of his father-in-law, apparently by 
a similar arrangement. In Pur. genealogies tne wives of all 
tfcwe four kings (and the mothers of their ^heirs) are called 
Prsadvati, - 

( 155 ) 

The substance of the tradition, apart from details and 
variations, mistakes and embellishments , may be put thus : 
A king of Yaya-tis' race, ruling at Pratisthana (and prob.= 
Ahamyati or Samyati Paurava, i.e., Yayatya), had a daughter 
Madhavi, also called Drsadvati, who, by some agreement of 
obscure motive and origin, 1 was jointly queen to four con- 
temporary and neighbouring kings 2 (viz., Haryasva of Ayodhya, 
Divodasa of Varanasa, Usmara- of the N.W., and Tisvamitra 
of Kanyakubja) , and who edified 3 and bore famous sons 
(namely, V&sunianas, Pratardana, Sm and Astaka) to four 
different families (viz., Aiksvaka, Kasi-Aila, Anava-Aila and 
Kausika-Aila-) , and at the same time secured for her father's 
race the * merit of perpetuation ' through daughter's son?. 5 
Later on she held a ' svayamvara ' afresh, 4 and finally went 
into exile with her last choice King Haryasva (ousted from 
his kingdom), 5 who was also the first; and their subsequent 
progeny became merged in the Yadava groups, 6 Madhavi is also 

1 The story is iold at LLilardsUa/t, cuui'l t^ illustrate tu the princes 

the evils of persisting in one's whim recklessly and of too mucn 
insistence on any one object ; apparently it is Galava's insistence 
on paying his guru's fee that is illustrated : but this clearly 
belongs to the subsequent brahmanical setting of the story $ 
originaljy the insistence exemplified before a Paurava court must 
have referred to an ancient Paurava court episode rather than 
a brahman teacher's fee; and the kernel of the story is in 
fact such an episode. The point of the illustrative story seems 
to be that by insisting on a dowry or bride-price of 800 horses 
of rare breed for his daughter/" the Paurava king of Pratltfaa,na 
had xo give her as common wife to four suitors, i and even then, 
the arrangement proving unsatisfactory, he had to offer her iii 
* svayamvara * again. 

* Prob. following the famous example of G-adhi of Kanyakubja in 
the preceding generation; cf. Av., V } 17, 11 15, ^vhere horses 
of precisely the same breed are a prized possession of kings. 
valued equally with a beloved ' rich-dowried * queen. 

t Who had other reasons also for a close combination viz., the 
common danger from the Haihayas (at this time allied, by 
marriage, with the Pratisthana court), 

2 For a Jataka parallelj vide infra. 

3 It is to be noted here that the Pracinvam-Ahamyati section of j:he 

Yayatya^Paurava dynasty evidently bejcame extinct at this point, 

'W JV i.J'CU-tifcJ A.\-TA* ,*,>! WbW> WV T *- J ^| ^ _ * 

branch. Thus the story about the * fall * of Yayati and his 
4 salvation ' through the fame of his daughter's sons had a his- 
torical foundation. 

4 Probablv because the first arrangement could not work well for long. 

5 His expulsion may well have been due to the other three kings. 

6 The g'lrasena section of the Yadavas (desc. from Krtavirv a ,_ conn. 

with Ahamyati by marriage, and thus with Madhavi and 
Haryas'va) " had just risen in the period contemplated by this 
episode; obviously the Haiiv. version has confused a real tradi- 
tion -re the affiliation of an exiled Aiksvaka family to the wfated 
gurasena-Yadavas, by identifying the ear ker fiurasenw ^ the 
later i^ce of Madhn, probably under the influence of the name 

( 156 ) 

said to have obtained a 'boon' from a f rsi' that after every 
connexion and child-birth she would regain her virginhood 
without prejudice to the next case, and she accordingly her- 
self suggests that polyandric arrangement ; and the four kings 
also are fully aware of what they and Madhavi were about, 
and show every sign of approval and delight; while their 
sons by her -are their heirs by preference. 

Such a remarkable tradition regarding famous ksatriya 
dynasties and heroes must have been well-known in the days 
of the Bharata war, and Vyasa as a Pauranika might be 
expected to refer to its precedent on the question of a ' sadha- 
rani ' wife for the Pandava princes. He does refer 1 to it ; 
only later mythical and edifying accretions have obscured this 
reference : the ' 4 Pane endropakhy ana ' ' is nothing but a garbled 
brahmanical account (with an admixture of folk-tale) 1 of this 
once famous and striking tradition about the Baurava princess 
MadhavI-Drsadvati and her four (or rather five) royal 

It is a noteworthy feature in the Madhavi-' Pancendra ' 
accounts that the polyandry described is not a : fraternal ' one : 
there is some amount of blood relationship between Madhavi' s 
several husbands no doubt, owing to common Aila descent and 
dynastic intermarriages, 2 and Madhavi herself is so related to 
them; but there is no immediate fraternal relationship 
between the four kings. So also the several * Indras ' (' Sivi, 

1 With this tale of one wife for five * Indras ' may be compared the 
still lingering folk legend of 1 Indram for 7 Indras (cf. 
a communicated note by G-rierson in J.R.A.S.). The P'uranlc basis 
of such legends may be traced to traditions like that of Nahusa 
courting ' India's ' queen "when he too became an ' Indra ' (Salya 
tells the story to Yudh. on the eve of the battle : Mbh- 
Vj 11 15.). The tradition of the common qneen of these four 
great kings 3 some of whom might well be called * Indras,' may 
also have bien one source of such a legend. It is noteworthy that 
Visvamitra's father was 'Indra' incarnate; and Siva and Pra- 
tardana were famous and powerful enough for the title; so also 
other Aila and Aiksvaka princes had actually become ' Indras/ 
Perhaps tlie ancient kings who were called or said to have be- 
come ' Indras/ only held or usurped the position of High Priest 
of the tribe or realm, in addition to that of King. Cf. the 
Devaraj and Dharrnaraj (or Dharma) of Bhutan, its High Priest 
and Chief Judge. So also Epic-Puranic tradition knows of 
1 Videha and 1 Iksvaku king as Devaraj (a), and 1 Vasistha 
with the same designation (vide Pargiter : AIHT. p. 342 for 
refs.)^ and Nahusa is called ' Devaraj * , (and equivalents) about 
24 times in Mbh. (V, and XIII) ; while Vidura and Yudhisthira 
were Dharma(raja)s. (Cf . also the current idiom, ' Indra-pata ' = 
passing aways of a great social leader). It is thus possible that 
the Pancendra and * Saptendra * legends are echoes of the times 
when High Priests (royal or otherwise) had oft^n wives in com- 
t mpn ( f< maharstni va Gautami " ; cf . n. l s p. 161). 
I This Is qnite apparent from the genealogies. 

( 157 ) 

Visvabhuj, etc.) are unconnected personages, the only com- 
munity being their suspended ' India '-hood or royalty. An- 
other feature is the initiative taken by the common wife. In 
the one case the brahman Galava plays an ill-fitting and 
almost uncalled for leading part, and in the other an advance 
is made by putting Siva in the same position. But it is quite 
evident that the rsi and the god are there to silence criticism 1 ; 
the chief share in arranging the polyandric connexions belongs 
to Madhavl and " Sri " 2 ; the former herself suggests such 
connexion and guarantees that no question of her s virginity * 
can be raised by the several husbands; the latter allures an 
4 Indra 9 into the ' cave * where four others have already been 
led to complete her quota., and paralyses her victims by her 
touch. A third feature is an indication that such a polyandric 
arrangement was incidental to times of great distress, 
expulsion from ' Indratva ' or lordship in one case, and that 
from their respective kingdoms in the case of the four con- 
temporary kings, owing to the famous Haihaya-Ya-dava 
invasions : evidently the connexion was intended to serve as 


the basis of a combination against the common danger. 

The parallels in the Paadava age are significant. Like 
Madhavl, Kunti is also granted a * boon ' or a * mantra ' 5 by a 
rsi, whereby she could, without detriment, summon any num- 
ber of notable persons C gods ') to her presence and bear 
children to them ; and after her first experiment she was 
granted a further boon ( if it was not already included in the 
first) that she would continue to be a virgin all the same. 2 
In connection with Draupadi's five consecutive marriages and 
consummations it is stated that every time she became a 
virgin afresh. 3 Satyavati, 2 steps before Kunti retained her 
"maidenhood" even after bearing a son to Parasa-ra by 
virtue of a similar * rsi ' boon. 4 Amongst the Yadavas, be- 
sides Eunti, Bhanumati, daughter of Bhanu a relative of 
Krsna, is given in marriage to Sahadeva-Pandava like an 
ordinary maiden, after her rape by Niloimbha, with whom 
she lived for a pretty long time before her rescue. 5 

I So also Qaruda is brought in and dismissed by Galava to supply him 
with divine sanction in his transactions, an improvement vp&n 

% gri=Madha-vi, in later mythological equations; this may be one of 
the starting points of the Sri' and PafiK^endra story. 

3 This was used by her co-wife Madri also : Mbh. I, 124. 

4 Mbh U31 (Kunti) : I. 67, 276S'74; $175 jKarna-sambh. } : I, 111, 

4385 ff ; $ 189 (Pandul : I, 122, 4748; 190 (Pandavotp. ) : I, 123 S 
4760 Of 569"(Bhagavadyana): V, 144; XT, 27, S 620 
(Sraddha. ); 789 (Putradars. ) : XV ? 2930; 547 JKarna) 
HI, 303307, etc. 

5 This was a ' rsi ' view quoted to Janamejaya : Mbh. 240 ( Vaivah. ) : 

T, 199 4eid). ' M f ^ 

6 Mbh 5 171 (Bhisina-Satyuv. ^ : T, 103; cf. 63. 

7 Kv. 149. 84718547. 

( 158 ) 

And Kunti's own sister Srutadeva, though married to 
Vrddhasannan of Karusa, Is stated to have been mother of 
Ekalavya, famed as Naisadi (and son of Hiranyadhanu) , 
having 'been brought up by the Nisadas near Dvaravati 1 : 
clearly, Srutadeva had a similar adventure to Kuntl's, and 
Ekalavya was her ' kanma * son, which however was no 
detraction from her ' maidenhood 9 or a bar to subsequent 
marriage. The frequent ascription in stories of restored 
maidenhood to * apsaras 'es (some of^ whom were real 
women) 3 after connexions with rsis or princes, is thus partly 
a reflex of actual conditions and opinions. This legal fiction 
of restored or continued maidenhood was evidently invented 
at a later period to justify undeniable cases of polyandry (and 
license) in the near past, or may have been coeval with that 
institution in its last days. Like Madhavi, again, Kunti her- 
self suggests to Pandu how she might become mother of 
children by other men^; arid like her and * Sri, 1 Draupadi 
captivates all the five brothers by gazing upon each one of them 
in lovg, when she is brought to the hut by Arjuna and Bhima. 4 
Subsequently, on the eve of the great battle, Krsna, the 
* sakha 5 of Draupadi had a secret conference with Kama, the 
' kanma 9 son of KuntT, in which he tried to win him over to 
the Pandava side, by promising that the covetable Draupadi 
will approach him also as wife when the 6th turn name. 5 Such 
a bait could not have been offered if Krsna's ( sakhi ' had not 
taken the initiative in the matter and expressed to him her 
willincrness 6 to extend the scope of her polyandry bv co-option. 
(The Pandavas it is said came to know the truth about Karna 
after his death 7 ; it may or may not be true ; but that presents 
no difficulty, as Panqu also did not know about the early 
amours of KuntT who persuades him that she was for the first 

1 Hariv. 35, 1937-8; together with V-Lyu : 96, 145 ii. (and corr. 

portions, i.e., ie Vasudeva's sisters, of Matsya, Brahmantla, 
Visnu, etc.); in B-rahmanda : III, 71, 189-'90 3 Ekalavya, " the 
child brought up by Nisadas, is ascribed to a nephew of 
Sratade^a; apparently her c kanina ' connection was with this 
near nephew. 

2 'Apsaras * status being ascribed to them owing to similarity of the 

names (like UrvasT 5} Menaka, Ghrtaci, etc.) wEicn were quite 
usual; e.g. the wives of Bandrasva and Pururavaa, or Visvamitra 
and' Bharadvaja, etc. 

3 She is no doubt first requested to bear children, bub the method for 

this is her own. Mbh. 189 (Pandu) : I, 122. 

4 Mbh. 236 (Svayamvara) : I, 192. 

5 After his embassy to Haatinapura, Krsna took Karna on his car 

and spoke to him of their being cousins and about "Drauuadi. etc. 
Mbk 569 (Bhagavadyana) : V } 140. 

6 The incidents at her ' svayamvara J and the * dytita ' partly explain 

how this willingness may have arisen. 

7 But Karna knew, at least from BMsma, Krsna ,and Knnti herself ; 

cf. also Mbh. 620 (graddh. ) : XI, 27;" 62 (Rajadh, ) XII. 
6: Yudh. had suspect&d it at the dice-match (from resemblance) : 

-rvTT, 1, 

( 159 ) 

and Sri stories Is found also in the case of Jatila-Grantaml, 
about a dozen steps later, where no relationship between the 
7 husbands is suggested, while in the- same connection 2 the 
10 husbands of Soma',s daughter are stated to have been 
brothers with a- common appellation. In Euntf s case (which. 
is as much one of ' niyoga ' as of polyandry), 3 some of the 
' husbands ' may have been related as half-brothers or cousins, 
but others were not. 4 In Draupadfs case also, it Is not purely 
& fraternal/ for Nakula and Sahadeva had no blood relation- 
ship with the other brothers ' at all, and were simply in the 
6 status ' of brothers ; the rest were but half-brothers. 
Madhavi's being the common wife of four kings did not/ prevent 
her sons by them from duly succeeding to their respective 
fathers' kingdoms (even by preference over other 1 sons, as with 
Astaka and Sivi) , or those kings from having other individual 
wives (as with Visvamitra and Usinara) and other sons by 
them. 5 So also in Draupadi's case, her sons by some of her 
husbands are recognized as * dayadas ' to them in- 
dividually, 6 and probably this was so in all cases, with 
the exception of Arjuna's son by Draupadi (being apparently 
born after Subhadra's son Abhimanyu) 7 ; and the Pandavas 
also have other wives individually, 8 though not without some 
opposition from Draupadi, 9 and other sons by them. Again, 
just as Madhavi is free to select a husband In the regular 
manner, even after her previous connexions, 10 so also Draupadi 
is asked by Duhsasana and Kama in the ' sabha. ' to select 

1 Mbh. 189 (Pandu) : I, 122 (latter part) and 123. 

2 Mbh. I, 196, 7266. 

3 For the ' nlyogas ' were not confined to one person, and Pandu all 

along lived with his two wives, exercising fall conjugal rights 
(at least subsequently), 

4 Vide infra, sec. on * niyoga/ re Kunti. 

5 Usinara married 4 other dtrs. of * raj arsis ', and their sons -were 

established in a number of Punjab principalities named 1 after 
them, the main line being continued by Sivi : cl. Brahma : 13. 
2024; Hariv. 31, 167479; Vayu : 99, 18; Brahmanija : III, 
74, 17 20; Visnu : IV, 18, 1. Re Visvamitra*s other wives and 
sons, cf. references to them In the Trisanku stories in all Pur. 
and the Kausika gotra accounts in the same, 

6 E.g. &atanika, Nakula's * dayada ' (often called Nakuiih) : cf. Mbh. 

VII, 1086. (It is to be noted that Nakula's son by an individual 
wife of his, Niramitra, is not his 4 dayada '). Prativindhya, her 
son by Yudhisthira, is apparently the latter's own ' dlyada ' 5 
cf . Brau.*s lament in the Sabha that she cannot bear the thought 
that Prati. should be Called a slave's son being the * raja-pntra ' 
(the King's or Yudh.'s heir). 

7 Mbh. 253 (Haranahar. ) I, 221. 

8 For these individual wives and their sons, vide : Vayu : 99, 240 *43; 

Matsya : 50, 51-57; Visnu : IV, 20, 11-12; and numerous refs. 
in MI detail in Mbh. itself to each of these wives and their sons. 

9 Mbh. 253 (Haranahar^) I, 221. 
10 Mbh. V, 120. 

( 16Q 

anew a husband from amongst the Kurus 1 (though the occasion 

for the request is a special circumstance) , and later on Jaya- 

dratha asks her to leave her five husbands and be his queen 2 ; 

Kicaka also wanted her : he did not know who she was, but 

knew that she was a maid-in-waiting with five husbands whose 

renegeance might fall upon him 3 ; the underlying idea appar* 

ently was that previous polyandric or irregular connexions 

(like those mentioned above) were no bar to subsequent 

regular marriage, The third feature of the Madhavi- 

Pancendra stories is also common to the cases of KuntI and 

Draupadi, particularly to the latter. On the continuity of 

Pandu's claim to the throne through sons raised by KuntI 

(who was a Yadava princess), on the securing of Pancala 

support and maintenance of fraternal unity amongst these 

' Pandavas/ turns the whole story of the Great Epic. 

* / *j **- 

For no instance of polyandry, however, is so much detail 
available as that of Draupadi; and an examination of these 
details should bring out what polyandry was like 4 in its last 
days amongst the ruling classes of the end' of thef Vedic period. 

Polyandry in some form .seems to have continued longer 

amongst certain priestly sections (as noted above) . When 

Utanka, a pupil of Veda (the ' purphita ' of Janainejaya III) 

is most calmly requested by the latter 's wife to take the place 

of her husband s.nd approach her for the sake of e virtue,' 5 H 

is evident that this was not a mere instance of laxity and 

adultery (which were common enough) , but & customary 

latitude allowed to the brahman wife, amounting to polyandry. 

So also, Uddalaka's 6 ' wife 9 is free to go with other ' brah- 

mans/ either of her own will, or in response to invitations, 

and this fully in accord with ' honoured rsi custom 7 ; and 

Svetaketu is her son by one of her * husband's ' pupils. 8 Such 

a state of affairs 9 would show that in priestly settlements and 

retreats, isolated from public city life, resident brahmans of 

1 Mbh. 304 (Armdyiita ) II, 7T7 (Dull. 's request) ; 300 (Dyuta ) II, 

71 (Karna's request). 

2 Mbh. 522 (Drau.-har.) III, 267. 
5 Mbh. 551 (Kic.) IV, 14 ff. 

These details are enough for a separate monograph ; it is interesting 
to follow the jealousies and conflicts of the co-husbands, and 
the changing favours of the common wife, or the legal and 
social position of the partis concerned so far as illustrated 
in different episodes. 

5 Mbh, I. 3. 

< Contemp. of Janamejaya III, cf. Mbh. I, 53a, 2047. 

n Mbh. 187b. (Pandu) "X, 122, 4724 '35 ; vide n. 2, p. 153. 

B Mbh. 635 (Kajaah. ) XII, 34, 1229. 

^ 6 steps further on (cf . Pargiter : AIHT. p ; 330) jSatyakama-jSbala 
is born of a woman who had connexions with a number of 
brahmaTis in one household (or establishment), so that the 
parentage of her famous son remained uncertain (Chand. Upan. 
TY. 4, 1-2), 

( 161 } 

a group often had a woman or women in common, 1 It is 
noteworthy that these two Instances refer to the JLngirasa, 
Kasyapa and Atreya groups, 2 otherwise noted for traces 3 of 
polyandry and laxity. 

For the Intervening period 4 between the latet 
and Buddhism, cases of polyandry are not known to the 
Puranic dynastic history. 5 But the great prevalence of 
metronymics in this age amongst the brahmans is suspicious, 
and cannot have been all due to polygamy, 6 for this was more 
or less general in various other earlier or later periods , and 
equally amongst the ruling classes. 7 This cro'p of metrony- 
mics 8 amongst the priesthood must have been therefore partly 
due/ to continued la-xity 9 and polyandry, in a proportion that 
cannot very well be determine'd. Buddhistic references to 
polyandry are not many, and these are mostly true echoes 
from the earlier Puranic traditions. Thus the story of Krsiia/s 
marrying the 5 Pandava princes 10 is told plainly and without 
fables, with the explanation that she was' a passionate girl who 
fell in^love with five youths at the same time, insisted on 
marrying them all (to which her father agreed rather reluctant- 
ly), and yet craved for a sixth consort: 11 quite in agreement 
with epic indications, again, Vidura the Kura (prince and) 
counsellor warns Arjuna's son' against having a wife in common 
with others, a calamitous thing for a householder ; yet it 
appears that his own sons had a common wife, on whom he 
relied for their guidance. 12 The story of Paficapapa, the 

$ In the orthodox ' sanghas ' of Buddha's time (i.e. brahmanical settle- 
ments) a few women were common to the whole congregation ; 
(one of -them accused Buddha of connexion with her); cf. the 
almost parallel practice in the late medieval Va-isnara "mathas 1 , 

2 Ud'dalaka's father Aruna was a Gautama ( Aruna- Anpavesi-Gautama) ; 

so also Uddalaka is stated fco have been an Angirasa (Matsya : 

196, 4. 6. 8) i he however founded an Atreya gotra (Matsya : 

197, 2); Veda, like Uddalaka, was in residence ^it.>> the Ka^yapa 
Dhaurnya; Veda was also an Arum (Varaha : 37, 7^. 

3 Vide ante and infra. 

4 Of 3 centuries, bet. 850 and 550 B. 0. 

5 Apparently owing to the concise character of the traditions for these 


6 As Keith supposes in his Ait. Aran. 

7 Vide infra, sec. r& polygamy. 

8 Later on in history there is a parallel prevalence of metronymics Jn 

.the Andhra inscriptions and coins; but snoh clear Bravidian 
character is not evident in the earlier case : though it is possible 
that some of these metronymics embody traces of matriarchy in 
the originally non-Aila brahman families, 

9 Of this several 'instances are known in contemporary literature. 
10 Cowell : Jatakas : V, 225 '27 ; 240; 243. 

12 basis of this particular may be either the epic tradition of her 
agreeing to marry Karna as her 6th Imsbancl. or that of her 
having a favourite eunuch attendant (Vr^annala. whom Arjuna 

U Cowell : Jatakas : VI. 126139. 

( 162 ) 

common wife of the princes Vaka and Pavarika 1 of Kasi and 
a neighbouring principality lower down on the Ganges, may 
however belong to the intervening pre-Buddhist period; and 
the introduction of Krsna's story to illustrate a contemporary 
statement, that a woman with even eight husbands (apparent- 
ly the limit reached by fraternal polyandry) yet longs for a 
ninth, 2 shows that the practice was not infrequent in Buddha's 
own time. Polyandry as an institution existed in well-known 
civilized states and communities in the Western sub-Hima- 
layan area, 3 in the post-Mauryan age. 4 It still survives in 
those outlying ' aryan ' tracts of country, 5 and amongst various 
Tibeto-Burman tribes on their border. 

1 Gowell : Jatakas : V, 236 239. This is a case of non-fraternal 
polyandry, the wife being shared in alternate weeks; (cf. the 
Pandava arrangements in the Epic) ; (the story adds that the 
queen co-opted a third husband to keep her company during her 
journeys between the two capitals). It is to be noted that all 
these instances belong to the Gangetic plains. 

Gowell : Jatakas : Y, 243, (so also, Vidura's warning against polyandry 
is applied to " all householders ", showing that the Jatakas knew 
it &s a nob very restricted custom). 

3 In Stri-rajya, Grama-nari (next to it), and Vahlika; the country 

"between and including Kumaon and N. Punjab. * Stri-rajya 
is known to Mbh., where its king is a candidate for the Kalinga 
king's daughter. 

4 Cf. Vats. Ka. But. II, 6, 4144; 39, 41 (with comm.)j also 
V, 6, 12 (re Strairajaka liar ems). 

5 E.g. in Rampur-Bashahr, Narkan<Ja (corr. to Nari-khancja, or 

Stri-rajya, Giamanari, etc.), and other districts around and 
beyond Simla, amongst the Kanwaris (who are popularly taken 
as = * kinnaras ' of literature) and other tribes; many of these are 
Aryan ethnically ; some are supposed to belong to the ' Khasa " 
race ; others are clearly Mongoloid. 


Xo case of 4 iNiyoga ' is definitely mentioned in the Epic- 
Pmanic tradition until about 41 steps below Mann. ; the next 
definite instances being at the 54th, 93rd, 94th, and 97th steps 
(with one not very long before the 93rd) - 1 This rarity in the 
earlier ages, and increasing number of cases later on. must 
partly have been due to gradual discouragement of polyandry 
and widow-remarriage 2 amongst certain sections of the ruling 
nobility, partly to increasing degeneracy of tie polygamous 
wealthy princes 3 , and partly to the growing pretensions of 
the priests. 4 The first circumstance would afford the scope for 
a specialised ' niyoga/ which would otherwise have been super- 
fluous; the second created necessities for dynastic continuity, 
whose urgency increase?, with the duration of those lines; the 
third developed a morbid esteem for introduction of sanctifying 
' rsi ' blood in the priest-ridden families. Indications of all 
these circumstances will be noted in the following account. 

No definite * niyogas/ again, are recorded of any other 
ruling family Besides the Eastern Anavas (Anga) , 'Aiksvakas 
(Hosala) and Pauravas (the Doab and ^Kuruksetra) ; while 

the brahman families expressly connecte'd with the practice 
are the G-autamas (Angirasas) and the Vasisthas, -with 
apparently the Kasyapas and the Atreyas, 5 all connected with' 
those regions and dynasties. There are a few -probable cases 
amongst the Paneala-s, Kanyakubjas and later Tadavas, 6 but 
hardly any traces amongst the Turvasas, Druhyus, "W. Anavas, 
Haihayas/ EMs 8 Vaisaleyas 8 and Vaidehas. 8 It would 

* The numbering is on the basis of Pargiter's comparative lists; the 
approximate general sequence would stand even if those number- 
ings have to *be altered later on. Of Mann's immediate descend- 
ants (within 3 steps?), Rathitara's wife is said to have undergone 
a ' niyoga J to an Angirasa. the resultant proerenv beinsr optionally 
known as Angirasas or * Ksat-opetah dviiatavah ' (Vis. TV, 2, 
2 f. and comm. on it; cf. Va. 88, 7; B<J. Iff. 63, 7: Hv. 11, 658); 
but acquisition of brahman clan name and of the above designation 
Is so frequent amonsrsfc Manva and Alia branch families (vide 
Pargiter AIHT.), and the alleged instance is so isolated, that it 
is more probable that the commentators' explanation arose from 
a var. lee. ', etc.,' in a Bd. test. 

2 As with the Hastdnapnra dynasty (cf. Bhima's -refusal to marry his 

brother's widows, and the singularity of the Pantfava polyandry). 

3 As with Vali, Vicitravirya, or Pandu. 

4 As with Angirasas and Vasisthas over various dynasties, 

5 Vide infra for the indications. 

6 Vide infra. , 

7 Except what- Is said in brahmanical stories about 'the ksetra^a 

ksatriyas amongst them after their defeat by the Bhrgns; vide 

infra. . , 

8 Though the Jtngirasas are directly connected with .the Vaisaleyas, 

and for a time with the Kasis, while the Vasisthas are similarly 
connected with the Vaidehas. 

( 164 ). 

seem as if the practice originated in the eastern kingdoms and 
spread westwards along with, the iiigirasa, Vasistha and other 
priestly groups, in the same way as Manva Brahmanism can 
be said to have spread to the Alias. 1 But the Kasis, Vaifialeyas 
and Vaidehas were as much eastern and priest-ridden as the 
'Angas and Kosalas ; the explanation may be the martial 
character of the two former, 2 and the absence of laxity in the 
latter. 3 So also the absence of the practice amongst Druhyus, 
Turvasas and W. AnaYas may be due to their having been 
virile fighting communities outside the Manva-Brahman 
influence; and though connected with the Bhrgus and Atreyas, 
the Haihaya-Tadavas were too strong and martial a race for 
priest domination, 4 and were vigorous, prolific polygamists, 
with a good deal of license in the sex-relations. 5 The main 
position, however, as stated above, is significant : the practice 
is associated with the Angirasas and Vasisthas (of Anga, Vai- 
, Kosula and Kuru-Pancala). 

The first 6 clear instance of the practice (that of 
Dirghatamas' sons by Vali's wives) 7 discloses several note- 
worthy features: There is no sign that it was regarde'd as 
unusual or novel. The brahman guest is already a privileged 
person, who is at once sent into the harem to have a pleasant 
time, 8 The previous history of Dirghatamas leaves no doubt 
as to how he used the privilege. It is after this that Vali 
commands his queen to obtain for him sons from Dirgha- 
tamas, who, like other solicited personages in later instances, 
agrees forthwith. Sudesna also readily assents, but afterwards 
pot liking connexion with ,a pur-blind man, substitutes a maid- 
In-waiting 1 (apparently a secondary co-wife, 'AtuSnari 2 , of the 

1 Vide Pargiter: p AIHT, pp. 303-14. 

2 About the Kasis, the mention of the Haihaya wars is enough; for 

the Vaisaleyas, vide the graphic account of Mark. Pur ana. 
^ Later on. in Asta\ T akra*s time, however, there were temptations at 
the Janaka court (Mbh. TIL 133V 

4 Of. their expulsion and oppression of these priests, leading to wars. 

5 As is evident from the Yadava dynastic accounts, and as no-ted 


6 Earlier legendary reference to ' ksetraja ' sons is very rare ; one 

such ( is ascribed to a king Svarastra on the Vipasa, driven out 
of his kingdom, whose queen had a son by a ' rsi,' who became 
the Tarmsa Manu (of uncertain chronological Dosition) ; vide 
Mark. Pur. 

1 The details that follow are given in full in : Mbh. 170 (Dirgh. ) : 
T, 104; fcf. XII, 342, 13182): 277 (JarasT 5 ) II, 21: (cf. II. 17, 
698; III:. 84, 8083; XIII, 7108; 7663; XII. 7593: also XJT. 
17961. Yisnu : IV, 18, 1-2 Brahma : 13, 28 ff ; Matsya : 48, 
23-24: 58-88; Brahmanda: IH, 74, 26-34: 36-99: HarW. 33, 
1683-90; Vayu: 99, 27-34; 35-99; 100-1. ' 

H For the much later post-Man ry an period also, Vats. Ka. Sut. refers 
to the practice of allowing brahmans free access to the king's 
wo?nen ; in G-auda specially ; does this show the eastern origin of 
this priestly influence? 

( 165 ) 

W. Anava family, and thus a cousin of the king], 
Dirghatamas then went on begetting one son after another on 
this Ausinari, and Ifc was not until the llth son had been 
born that the substitution was made known to Vail, as he 
now claimed them from Dirghatamas; from the details it is 
clear that Dirghatamas was allowed to live for ail these years 

^ * ^ 

within the palace in the same relation to the whole harem as 
the king himself 3 , but all the while he was living specially 
with Ausmari 4 ; the claim after the llth birth is significant; 
probably the eldest son having completed his i"2th year had to 
be definitely 4 affiliated J( in view of usual ceremonials. After 
the disclosure, Sudesna was sent for * niyoga ' once again, 
and this time there was no difficulty, the prolific brahman 
having apparently made the harem all 1 his own. After 
Sndesna had borne 5 (or probably 6) 5 SOBS by * niyoga/ 
Dirghatamas got full rights over Ausmaii and continued to 
live with her separately, begetting other children on her, as 
well as on other women (who may well have been inmates of 
Vali's seraglio like Ausinan). The scene of all this is placed 
in Girivraja 6 , where Dirghatamas 1 own family became settled, 
while the 5 ksetraja princes settled in 5 different provinces of 
the original kingdom, which seems to have included a large 
part of Bengal, Bihar and Orissa, with Girivraja as a chief 
centre ; and later on the 5 princes used to pay visits to their 
real father In his retreat at Girivraja. Three things are most 
striking In this common Epic-Puranic tradition : the revolting 
license of the (Angirasa) priest, the laxity of harem life, 
the utterly priest-ridden 7 and incapable type of king. 8 All this 

1 ' Sudra Ausinari ' may have been her full name ; one of the Panrava 

King Randrasva's daughters was named gudra. (Possibly Sudra 
was also the name of Vidura's mother). 

2 Of. the parallel case of Ambika the chief queen similarly substituting 

a ' maid ' -who is also a co-wife and apparently a princess. Cf . 
also the Puranic legend of Surenu's substitute, which shows a 
similar custom'. Austaari, shortly ' Ansi,' is a better source for 
the metronomic Ansija, than Us'ij, which is otherwise ^ unknown 
as a feminine name; the epic version is clearly in the right here. 
Princesses in the harem suffering frequent changes of status, 
owing to royal or their own freaks, was very common all along; 
cf. Buddhistic references to pre-Bharata and post-Bharata court 
stories, and Yedic references re 'parivrkti/ etc. 

3 Cf . the chosen brahman agent ' living with ' Saradanijayan? till 5 sons 

are born to her; vide infra. 

4 So also Krsna-Dvarpayana was particularly pleased with Yidura s 

mother. . , _ . , , - 

5 Including Anapana. (It is prob. better to read soparasca in 

the text for c saparadha/ etc. ; prob. also the real name was 
Annapana=' food-protector '; cf. gaJi-vahana). 

6 The epic tradition is very clear and consistent with regard to this 

Ideation of the episode; so also the Puranic : e.g. Mat. 48, 84-btf ; 
Va. 99, 37-99; etc. . ._ 

7 Also shown y the humanistic 'economics ^J^J^^ 

8 Cf. " he was born when the race had dwindled : Matsya: 48, 


( 166 ) 

cannot have developed in a day ; the sort of * niyoga * as 
exemplified amongst the E. Anavas and Angirasas, therefore, 
must have been an accepted and established practice long 
before the 41st step from Manu; the E. Anavas had not 
separated from their more vigorous kinsmen, the N.W. Anavas, 
for more than a century and a half, 1 and their rapid 
degeneration implies some pre-existing tradition of harem life 
and priest-influence in the land of their settlement, already 
peopled by the Saudyumna ' and Manva races. 2 

The 'niyoga 3 of Madayanti, Ivaknasapada's queen, 3 also, 
discloses somewhat similar features, the main difference 
with the previous case being that Vali takes the practice for 
granted and is glad to employ it, while Saudasa-Aiksvaka is an 
unwilling victim of it. It would almost appear from the 
details given about Saudasa's persecution of Vasisthas and 
ingirasas, 4 the curse of the injured Angirasi, and his final 
reconciliation with * Vasistha/ that his queen was part of 
the price he paid for his restoration (which was assured when 
Saudasa had actually solicited Vasistha to beget a son on 
Madayanti) , and that there was an element of retaliation and 
humiliation involved in the whole affair. Madaya-nti seems 
to have come into touch with "Vasistha 5 even before the 

* rdyoga/ while the king was in exile 6 , ; and when on return he 
approaches her, she dissuades him from his desire of begetting 
a son himself, and then Vasistha is asked to visit the queen, 
with whom he remains till she is with child. 7 On the whole 
what is an accomplished fact in the earlier case, is shown in 
the process of being completed, or reasserted after temporary 

1 About 14 steps before this, branching off from Maiamanas, under 

Usanara and Titiksu. 

2 As the dynastic accounts clearly show: for details,, vide Pargiter : 
" AIET., Chaps. XXIV and XXV. 

3 The following details are given in fnll in : Mbh. I. 182, 6888 ff : III. 

218, 14128, etc.; I, 122, 4737; 177, 6768; 6791; (cf. 176-177) 

W\&J^> ^ 8604 > V5 y n: 88 > 176 ff * Brahmanda: III, 
63, 177 ffj Vi 5 nu: IV, 4, 19-38; etc. 

4 At the instigation of the Aila < Visvamitra.' 

5 So also, while Trisanku remained in exile for 12 years, ' Vasistha ' 

protected the royal harem and the kingdom, and the latter 'resented 
it very much; (in all Pnr.). 

6 One account makes her accompany her husband in his frenzied 

wanderings ; it is not however clear whether the exile had begun 

? And seems to be connected with her later on 

( 167 ) 

protest and check. 1 It is notable that while nothing is said 
regarding Yali's merit in lending his wife (or wives) to 
Dlrghatamas (though he is generally lauded as a pious king), - 
Saudasa by * giving his dearly loved queen to Yasistha * (not 
simply ' raising a ksetraja son ') is declared to h&ve 
' attained heaven together with that wife \: a befitting 
praise for a fresh or repentant 1 convert to the system. 

The nest group of clear cases of * niyoga ' (of 
Vicitravirya'0 wives, Pandu's wives and Uddalaka's wife) are 
different in features from the above two. They do not show 
that domination of the king and the harem by the priest in 
the presence; of the king himself. The court life is equally 
lax and degenerate 3 if not more so; and the continuance of 
the dynasty is equally a necessity; but that end is achieved 
through relatives or equivalents of relatives, and nofc 
through an unconnected priest as such. 2 In these ' niyogas ' 
therefore, another element is present, the rights of kinship ; 
the practice in this form is a corrollary to and an off-shoot 
from ' group ' or fraternal polyandry, while the form 
typified by Saudasa's and Vali's cases is derived from 
ascendancy and pretensions of the priesthood. 3 This, 
however, was still present : Kunti is referred to an (apparently 
not much) earlier definite instance 4 of the * niyoga ' of a 
Ksatriya 5 wife, Saradandayam 6 , who, at her husband's request, 
came out prepared into the public square 7 and selected and 
solicited a suitable brahman from amongst the passers-by as 
the agent, and had successivly three sons by him alter due 
ceremonials. Pandu mentions brahmans "amongst others 
as suitable agents Eunti might think of. Bhisma, citing in 
full the instances of the Bhrgns and Dirghatamas, 
recommends a * rsi ' agent; to Satyavati when she presents to 

1 The latter is more likely, as tlie Angirasas and ysithas were long 

since intimately connected with the Manvas,, and bad other con- 
flicts Tvith them before. 

2 The Paurava princes had many struggles with the brahmans, and 

were only partially and for short periods under their sway. 

3 Of, 'the claim in AV. : the brahman has -lights over every wife of 

every other man ; cf . a revolting example in the Epijc story of 
OghayatI 1 (a Saryata-Tadava princess and wife of a Nila (Panrava) 
prince of Maliismatij settled in Kuruk'setra), who was enjoyed 
by a brahman in her gratified husband's presence, by right : 
Mbh. 720. b. (Sudarsanop. ) : XIII, 2, 122 if. 

4 Mbh. I, 4677-9. 

5 * Vi'rapatni ' ; but ' Vira * may be the husband's name. 

6 The name looks brahmanic ; she may have been a brahman Saradancla's 

daughter married to a ksatriya or a prince, not an unusual 
thing. [Q. Is she the same as Sarakanta's dtr. (apparently a 
corr. reading), w. of Andhaka Mahabhoja (Padma. V, 13. 45) ? 
an that cae Kunti -was aptly ref erred to her example.] 

7 For an Aikakava parallel (of somewhat- later period perhaps), vide 

n. 10, p. 220. 
8 Mbh, I, 4680, 

( 168 ) 

him the case of his xvidowed sister-in-law. 1 Bui in these 
5er instances, Pandu also mentioned his brothers inends 
and *ood men ' (equal or superior to him) as his substitutes, 2 
_ a nd%he first proposal of SatyavatI was ' myoga, or re- 
marriage of Ms sisters-in-law with Bhlpna (the elder brother) 
himself while her last and finally accepted proposal was their 

'niyocra' to her own illegitimate son Ktsna-Dvaipayana 

(equally an elder brother) 3 ; it was accidental that he was a 
'rsi-' he was expressly selected for being an elder brother 

(on the mother's side)*, though Bhisma is made to approve of 

it doubly because he was a ' rsi.' 5 

The ' niyoga ' of Uddalaka's wife to his disciple 6 belongs to 
about the same age 7 ; but though there is one common point, 
in the selection, -as agent, of a pereon who is almost a member 
of the same family (in theory, if not by blood, which was 
sometimes the case), it stands on a somewhat different 
footing. As noted already, ' niyoga ' of this type is but a 
form of the general license that prevailed amongst brahmaii 
settlements (which may have been connected with a sort of 
' group ' polyandry). The brahman disciple indeed was 
often regarded by the preceptor's wife as being in the status 
of her husband (as shown by Veda's wife's request to 
Utanka), 8 in spite of all the denunciations and prohibitions 
of the (later) brahman law-givers, which only show what 
actual conditions often were. If for instance Utanka had 
consented (as others like him evidently did), Veda would have 
had a ' ksetraja, ' son by a ' niyoga ' arranged independently 
by his wife, becamse lie was absent. It is noteworthy that 
Uddalakanfs is the first, and probably thet only recorded^ 
example of * niyoga ' of a brahman woman 9 while brahman! 

1 Mbh. 169-71 : 1, 103 ff. 

a Mbh. I, 4671-80. 

3 Called ' devara ' in the text ; this word therefore applied to all the 
"brothers of a husband; so also Ambika understands Bhi?ma by 
* yonr devarV (Prob. the original meaning of * devr * is a person 
with whom 'dalliance or amour ' is permissible even in the 
married state.) 

4 The one on the father's side declining. 

5 It is to be noted that the Vaithas (to which family Krna- 

Dvaipayana belonged) had become connected with the Pauravas 
from Sarrwaraiia's days. 
8 Mbh. XII, 34, 1229. 

7 Three or four steps lower. 

8 It is to be noted that Veda was an Angirasa (Gautama), being an 

Aruni; vide ante. 

t Even this can hardly be called a 'niyoga,' for Uddalaka's wife was 
certainly not restricted to one husband, and probably the ascrip- 
tion of a ' niyoga * may be nothing more than giving a better 
name to some acknowledged connexion with ^ disciple. (In Mbh. 
XII, 34, 1229, tile justification cf her case la that connexion 
with a ' gurnpafcni * is no sin if the result is for the benefit of 
tbe ' gum '). 

( 169 ) 

polyandry is much earlier and more freqoeut; continued 
laxity and polyandrous relations probably accounts for tills 
comparative rarity of * niyoga ' among them ; besides, lineal 
and engenic continuity did not concern the brahmans much. 1 

In the first of the two Kaurava cases, the ' traternal * 
character of the L niyoga * becomes further clear from the 
detailed description of it 2 ; Satyavatl persuades her elder 
daughter-in-law to the * niyoga/ and tells her that a brother 
of her husband will approach her at night in her own bed- 
chamber; and Ambika began to think of Bhisma and other 
elders of the Euro family (evidently the sons of Yahlika, 
elder brother of Santanu, who were almost always resident 
at the Hastinapura court, though they had inherited their 
father's maternal uncle's kingdom in the Punjab) 5 ; she seems 
to have been taken aback when her expectations did not come 
true, and she found in her room Krsna-Dvaipayana instead , 
-evidently, and naturally, hitherto unknown to her 4 ; and 

u< * nJ J 

afterwards she declined to have another SOB by him. 
Satyavatl repeated the same instructions to her second 
daughter-in-law Ambalika, and she too behaved almost in the 
same way. The attitude of Ambika 's ' maid J (or co-wife) 
to Krsna-Dvaipayana was entirely different : she apparently 
had no preference for Bhisma like the other two princesses, 
whom he had seven years back abducted at their 

' sva-vamvara, 1 and then made them over to his brother. 5 

K * 

All this is the direct evidence of the Epic as it stands 
now; but there are many indications that some of the 
original relationships of the KuraJPandava tradition have been 
revised 6 to suit later tendencies towards edification and 
mythological explanations; this (latter) is more apparent in 
the case of Strati's * niyoga ' than in the earlier one, for 
obvious reasons. The * revision ' in the earlier case seems 

1 This affected brahmanical genealogies as much as any other eircum- 

stanjce; cf, Pargiter : AlHT. pp. 184-85; Vedic evidence also is 
very clear on this point. 

2 Mbh. I, 106 ( 171 : Vic.-Sntotp.). 

3 Cf. Mbh. V, 149, 6055-' 67, etc., for the arrangement; they were yet 

called ' feanrava-dayada'sj ete., e.g. VIII, 5 f 106, etc. 

4 His parentage was made known to Bhisma, for instance, only & ew 

days before the ' niyoga/ All that is said about his being an 
old * rsi * with matted locks, etc., is clearly wrong and late, as 
he was only slightly older than "VIcitravirya 3 who had died at 
about 23. 

5 Mbh. V, 173. . 

6 The probabilities suggested in the following paragraphs have the 

great advantage of clearing up a good ^deal of the tangle of 
dynastlte relationships and resultant clairns which led *o th 
great battle; and the parts taken by various people in the epic 
events-series become more intelligible. 

( 170 ) 

to have consisted in ascribing to one person Krsna- 
Dvaipayana all the three ' niyogas/ because he ^ was in a later 
age regarded as a .supernatural ' rsi/ an incarnation of 
Visnu, 1 and because Bhisma was becoming more and more 
an" idealized type. 3 Kuntl's case itself proves that it was quite 
usual to solicit several suitable persons to perform the 
' niyoga * \ thus Kunti underwent three different ' niyogas/ 
and her co-wife Madri two (legally equivalent to one) , 3 and 
more invitations were thought of for both. The reference to 
Satyavati's first thought of causing Bhisma to continue the 
line, and to Ambika's expectation of Bhisma and other elder 
brothers (or cousins) of her husband as agents, is a plain indi- 
cation of what must have been the original procedure : the 
last king's next of kin were invited to raise up offspring 
Hastinapura) , Bhisma and Krsna-Dvaipayana ; and the 
three ksetraja Kuru princes may well have been begotten 
by these three relatives on the three wives 6 of Vicitravlrya, 
quite in accordance with normal custom. So also, Bhisma 
and Vahlika are consistently called direct ' grandfathers ' of 
the Dhartarastras and Pan<Javas, and ' fathers ' of Pandu, 
Dhrtarastra (and Yidura), equally with Krsna-Dvaipayana ; 
and on the whole the parental connexion is asserted specially 
between Bhisma and Pan.du (with Pandavas), between 
"Vahlika and Dhrtarastra (with Dhartarastras) and between 
Krsna-Dvaipayaria and Vidura; and this in spite of the re- 
iterated derivation of all of them from the last progenitor. 

The indications ( regarding the relatives who may have 
taken part in the ' niyoga ' of Panda's wives are by no means 

1 E.g. in Mbh. XII, 350 (first part). 

2 That he was originally not so, is revealed by Sisupala's denunciation 

and several other episodes connected with Sis career 5 on the 
whole, however, he represents a better type than most other epic 

5 Vide infra. 

* In fact Vahlika and his sons and grandsons represent the Kauxava 
dynasty proper; the rest are questionable grafts (hence the justi- 
fication of their continued use of ' Kuru ' titles). 

5 Even so, of Kaurava blood, for Satyavati was of Vasu's line (which 

is the real basis of the BhIma-Matsya compact). 

6 Who may have had a choice between -those three; in that case the 

story of Ambika and Ambalika's dislike of Vyasa would originate 
from their actual preference of Vahlika and Bhisma and rejection 
of ^ Vyasa, who thus fell to the lot of .the 3rd wife. (Of. the 
rejection by Sudena of Dlrghatamas who falls to the lot of an 
inferior queen at first, and the choice by Kunti herself of the 
persons for * niyoga '). 

( 171 ) 

so clear. Of course all those fables about invocation of gods 
are only fables intended to obscure the actual partis, so as 
to remove the ordinary or the discreditable features of the 
life history of persons made into brahmanical heroes, and 
impart them something of mystic and divine glamour. 1 
Stray references to actual origins, discoverable for the next 
previous generation, are almost absent for the Pandavas, 
having been evidently laboriously modified or weeded out 
through centuries. Still some details and incidental notices 
suggest possibilities, much sounder than those fables. For 
one thing it is to be noted that the Epic and Puranic (and 
even Vedic) tradition knows of lots of other kings 2 who are 
said to have been born from Dharma, Maruts, or Incba (or 
the Asvinsj, being their ' portions/* gifts, sons, or incarnations; 
and these ascriptions to the Panaavas are by no means a 
special mythological conception. 5 Thus the detailed 
development of these _ common expressions and notions in 
connection with the Pandava, origins only shows that the 
details (of ' niyoga ' etc.). were already there in the original 
account, but tiiat there were some i'acte in it (like similar 
forms and imports of names) which easily led to those details 
being ascribed to some of the usual divinities whose 
prototypes or essences Mngs were popularly held to be. 4 If 
at the same time it is remembered that in the same family 
and in the next previous generation the 4 niyogas ' (whicli 
were the first sources of the Pandavas etc.) were performed 
by one or more near relatives, the conclusion becomes 
irresistible^ that ' Dharma, 3 ' Maruta/ ' Puruhuta/ and 
Asvinau ' (with' their .various synonyms) 5 stand for real 
kinsmen or relatives of Pandu whose names happened to be 


1 The prosess is not unique In Indian history; champions caught hold 

of by brahmans were translated into myths even in medieval ages, 
as when the barbarian Gurjadas, etc., were declared to have sprung 
from Agni at a sacrifice (a fable ascribed to other contexts in 

2 E.g. Drupada born of Maruts : Mbh. I, 67, 2715; so also Satyaki : I, 

67, 2714 ; and Virata : ibid 2717 (in his case prob. Marata -was 
also his real patronymic; vide infra.): so again., Pan$ti : XV, 
Slit. 851; Krtavarman: XVIII, 5u, 159; Mandhatr : ' XII, 29, 
974; cf. Maruts associated with Uarutta (in Pur. *). Similarly 
other kings are said to be born of other gods, 

3 Hence it is no^ ground for holding that the Panaavas were an un- 

known _ foreign mountain clan who invaded as usurpers, -etc. 

4 Such transitions from facts to myths, from names to fables, is a very 

common feature of early tradition, whether Ptnranlc or Vedic. 

5 Thus Ar juna as * son of Vasava ' suggests that a related * Vasavs * 

prince (desc, from Vasu of Cedi) may have begotten Arjuna. 
But as ' Marnta ' stands for one * Vasava * prince (vide infra), 
and as * Puruhiita' occurs in a genealogical verse, Purojtt (Kunti- 
bhoja) is probably the more likely agent (vide infra). 

( 172 ) 

such as to make later identification with those divine beings 
but an easy step involving no grea<t textual change. 1 In fact 
on a detailed examination of the Epic and Puramc relation- 
ships of the Kuru-Pandava and connected dynasties, it 
becomes sufficiently clear that ' Dharma ' might well 
represent none other than Vidura-' Dharma ' (younger step- 
brother of Pandu) , 2 * Maruta/ a Vasu-ite Cedi-Matsya 
prince of the same name (ruling over a people of also the same 
name;, being a cousin of Pandu through SatyavatI (and 
probably the predecessor of Virata), 3 ' Puruhuta/ Kunti's 
foster-father Kuntibhoja's son Purujit, being her cousin, 4 
and the ' Asvinau/ eithelr one or two (jointly ruling or 
actually twin) princes of the ' Asvapati ' or Madra family (to 
which Madii belonged). 5 Some of these would thus be relatives 
of Pandu through marriage; so also iu the earlier k niyoga,' a 
relative in the female line is selected as equally good with 
one in the direct male line. 6 Thus the meaning of Pandu 's 
recommendation of brothers and suitable relatives as his 
srbstitutes is elucidated; indeed, when ' Dharma ' is invited 
by KuBti, he hurries on a chariot to visit her, and accosts her 
smiling, and she too smiles knowingly, and solicits him for 

1 So also it seems very likely, from the great laxity in Yadava dynastic 
moral 8, that Kunti's * kamna ' son. by * Bhan.u/ etc., \vas really 
by her kinsman Bhanu-jt'adava (whose abducted daughter Sahadeva 
married later on) (vide infra). 

3 Apparently Vidura was also called l Dharma ' ; he may .have been a 
' judicial officer ' at the Kuru court (just as he filled other offices) \ 
cf. the Ani-Mandavya, story : Mbh. 1, 106, 4302 ; 107, 4506 ; XV, 
28, 752.754; ci also the Mbh. account of 'Vidura (=I)harma) 
entering Yudhisthira's body after death.' From all this it would 
seem that Vidura-' Dharma ' as judge had punished a brahman buy 
for cruelty, and the brahmans out of spite spun stories based on 
his parentage; it also becomes clear how Yudhisthira could be 
said to have been begotten by ' Dharma ' (so that ' being so 
begotten no blame could attach ' to KuntI and Pandlu). 

3 Maruta is the name of a Vasu-ite (Kaurava) prince or his line, in the 

Puranas (C'edi genealogy) his line was the same as the Mavellaka 
or Matsya line (vide Pargiter AIHT. 118-19) ; Marutah are a 
peo'pl-e amongst Yudhisthiras allies, and so are Mavellakas, 
prob. the same as Matsyas. Cf. Mbh. VI, 2083 (Bom. Edn.) -, etc. 

4 The genealogical verse is Mbh. I 3 126, 4921 : ' Puruhutad ayam 

jajfie Kuntyam eva Dhanairjayah,' for which may be read : ' Puru- 

jito hy ayam ' or ' Purujitas tvayam ' He Purujit 

(Kuntivardhana), cf. Mbh. II. 14, 581; V. 172, 5922; VI, 25.0. 
834; VII, 23.0., 995; 25, 1103; VIII, 6, 172. 

5 Many other Madra and Kekaya princes were called Asra-patis. The 

Mbh. account of Madii's m. makes Madra =Vahlika family; in 
that case the relationship was double, an earlier Vahlika prince 
having had a share in the Vaicitravirya nijrogas- If Madra- 
Kekaya, then also there was additional relationship through Kunti'a 
sister and her 5 sons 3 Kekaya prinjces; 2 other r contemp, Kekaya 
princes (prob. twins) were Vinda and Anuvinda. 
S Mbh. 168-171 (Bhi$ ASat.) : I, 103 ff. 

( 173 ). 

a sou 1 : there is little of the mythical here, and the description 
would apply equally to Vidura-Dharma as a fi didMsu devr ' 
and to Bointi as a ' devr-kama.' 

Some other features are also disclosed bv the Hum- 


Pandu cases. First, as to the sources from which the 
practice was adopted : It is not necessary to suppose that, 
along with Pandava polyandry, the Epic c niyoga ' implies 
Himalayan and non- Aryan origins, and that the Pandavas 
therefore were uncouth foreigners. 2 It has already been 
shown that polyandry was known amongst brahmaxi and 
Ksatriya families of the Gangetic plains in several earlier and 
later 3 periods; so also, there were earier and later cases of 
8 niyoga ' (as noted above) in various other families and 
parts of the country, even in the same dynasty. It is not 
a special case here. The only connections discoverable with 
Himalayan regions are, firstly, the accident of Pandu 's living 
on Mt. Satasrnga 4 at the time of the ' niyoga ' (where the 
reason given for that retirement is sufficient from a common- 
sense point of view), secondly, the statement 5 that on the 
birth of the 3 ksetraja sons of VicitravTrya, the Uttara and 
Daksina-Kurus vied with one ajiother (which would rather 
point to the Vaicitravliyas being foreigners, if at all), and 
thirdly, the inclusion of some Himalayan tracts within 
Yndhisth'ira's dominions (which fact is later than the cases in 
question). 6 The Kuru-Pandavas may have, more probably 

1 Mbh. I. 122 ff . ; the next ' agent ' also approaches her smilingly, 

and Kunti is liere * salajja * as well as smiling; she would 
naturally have been less familiar with the Maruta-ilatsya cousin 
than with Vidura. 

2 A view repeated in many very recent works. 

3 E.g. in Buddhistic references. 

4 Such retirements for various reasons (real or alleged) of one of the 

brothers are not uncommon in the dynastic tradition : e.g. Yati, 
Jyamagha. Devapi, etc. ; probably Pandu was actually exiled 
with his wives by Dhrtarastra, in the same way as the Pantjavas 
were ousted by Duryodhana. 

5 This might be taken to indicate that ( niyogas ' were commoner In 

Uttara-Kuru (where another primitive custom, that of sister- 
marriage, sp. bet. twins, was an established one, ace. to Mark. 
Pur.; also unrestricted polyandry^ ace. to Pandu in the Epic.}. 
g A l&rge portion of the Southern Himalayan region was, jfrom much 
earlier times, under the Aiksvaka (Manva) and W. Anava king- 
doms. That part of it where Pandu went to live, corresponds to 
modern Gadhwal, Sirmur and Kanawar (der. popularly from 
4 Kinnara '), where polyandry is still recognized, Tfee^ Kinnaras 
( Kanwaris) are named amongst the real Human Gandharvas in 
the Epic; and Draupadi' escapes detection by giving out that 
her five husbands are * gandharvas ' ; probably eyen in the Epic 
period these Kanwari ' gandharvas * had the same institution of 
polyandry as now. These points however do not prove the 
Himalayan origin of the Pandava polyandry and * niyoga- ; 
they only show that while polyandry was becoming rare in the 
plains it still prevailed in outlying hill districts; the surround- 
incr polyandry at Satasrnga may however have encouraged tlw 
' five ' * ( niyogas * of Pandu's wives. 

( 174 ) 

adopted these practices of polyandry and ' niyoga ' (if they 
were not known to them before, which is not very likely) 
from the Vasisthas, Angirasas and Kasyapas, they came so 
in contact with, specially from the days of Bharata 1 onwards. 
Bhisma relies on Angirasa precedent of ' niyoga,' Pandu on 
that as well as Vasistha; a Vasistha takes part in the 
Vaicitravirya ' niyogas/ and various * rsis ' justify the birth of 
Pandia's sons ; a Easyap'a priest (Dhaumya) , and that 
Vasistha, advise, sanctify and legalise the Pandava polyandry, 
without objections and armed with precedents. 2 

Then as to the number of ' niyogas ' permissible": Hunt! 
is made to say that connexion with the fourth man besides the 
husband makes the wife a courtesan, 3 and therefore she 
refused to undergo a fourth ' niyoga *i ; but she had already 
exceeded that limit, for actually 1 she had had 4 connexions 
excluding Pandu, and she had "4 sons by different fathers who 
all were or came to be regarded as Pandu' s sons. 5 Pandu 
indeed wants to have quite a numb'er of such sons ; and he 
had a precedent for it ; Vali had practically 17 sons raised on 
his two wives, 11 on one and 6 on the other ; Saradandayani 
raised 3 sons b"y ' niyoga ', and if Vyusitas'va's case 6 is "really 
one of ' niyoera , ' the number permitted is 7 ; on the other 
hand Ambika had one actual * niyoga 9 and another proposal, 
and Madayanti only one. On the whole therefore the number 
of ' niyogas ' was not restricted by any standing rule ; nor 
was its nature regulated by austere injunctions found in later 
codes: for in almost all the traditional cases of * niyoga,' the 
partis take to it with an evident element of initiative 7 and 
choice, personal feelings and attractions 8 ; and very often the 
wife is allowed to woo and choose afresh one or more persons 
of her own Accord, in view of such temporary unions, whose 
duration might be extende'd considerably (from one night to 
twelve years or more) . 9 

1 Vide infra, the prob. case of ' niyoga ' with Bbarata. 

2 The Buddhist echoes of Puranic traditions indeed suggest (vide ante) 

that polyandry was more frequent amongst the Kurns than 
appears from the Epic, and was known to Paficalas, Koalas, 
Kasis, and peoples further down the Ganges. 

3 Of. the same view in Vats. Xa. Sut. quoting Pancala-Babhravya (at 

tewt cir 600, B.C.)- It is interesting that the dictum is given 
a rancala origin. 

4 The fact is kept secret from Pandu. 

5 Re Karna, cf. " Pandu's son by" ; dharma ' " : Mbh. V, 140, 4734: 

141, 4756. ' 

6 "^ide infra. 

7 S; s * MEdri clearl JJ ^uatl and Madayanti partly. 

8 a xtT 6 ? ^ St features ^y be found m all the traditional cases. 

9 i>o that niyoga ' often verged on ' co-option > of a husband, or 

biandiy, or polyandry ; thus Lfadayanti is aaid to have been 
given to Vasistha, though she remained Sfudasa's queen all 
S?li! ' Allsillar! a11 ^ Saradandayani (ante) < living ' (for long) 

with their apportioned or selected ' agents/. 

( 175 ) 

Connected with this is another feature of these * niyogas * 
of tradition : continued (political or social) connection with, 
and (parental or kindly) interest in, the ' ksetraja * children, 
on the part of the real progenitors. That of Krsna-DYaipayana 
is inseparable from the Epic events ; and so is that of Vahlika 
and Bhisma, if they may be given the status that is almost 
plainly indicated for them; so also, of the probable originals 
of ' Dharma,' * Maruta * and * Pnrnhtita ' (who as such 
divinities continue to show parental concern), Vidnra*s 
interest in Yudhisthira is particularly strong, and the only 
three allies (apart from relations by marriage) that the 
Pandavas had at first, were Vidura, Virata the Vasu-ite prince 
of Matsya, and Purujit the Kuntibhoja. The fatherhoods of 
Pandu and VicitravTrya are but faintly asserted behind the 
prominent repetitions of the real relationships, which axe 
openly and proudly acknowledged before all and in the 
presence of all the partis concerned ; and it is the mothers who 
are more prominent than the putative faffiers (naturally 
enough 1 , though against the presumption of later law-books) , ancl 
who give their sons their better-known and more frequently 
used metronymics, while the s actual J patronymics are also 
applied. 1 In the Vali and Saudasa cases also there is the 
same connexion and interest : * Va^istha * remains in toncli 
with both the mother and the son, and his own son * Parafora * 
becomes that other son's guardian after Saudasa's. 'death 2 ; 
PO also Dirghatamas was often visited by the Valeya princes, 
and h'*s other descendants (by another wife of Vali) 3 continued 
for generations to protect and favour the- related Anava 
princes. 4 Only in Saradandayani's case the re&l father passes 
out of view ; but probably not so in the Uddalaka case, if Kalto- 
da 5 is the disciple referred to, as Kahoda lived all along with 
him, and also married Uddalaka's daughter, and her son 
Astavakra and Svetaketu were closely connected. 

* * ** 

Xnoffier noticeable feature of the * niyogas s of^ tradition 
is that the majority of them are performed in the lifetime of 
the husb'an'd, 6 whose 'disabilities as sucH are not always 

1 E.g. 'Dhrta. son of Vyasa 1 : Mbh, I. 1, 95: 60, 2213: 63, 2441; 

" 67, 2719, etc.; cf. 171 (Vic.-sutotp.) ; VI, 504, etc. ; 'Dhrta. , 
son of Ambika* ": 31 mentions; cf. Sor. Indes, pp. ^0-52: 
also * Ambik-eya ' : 8 mentions ; "bnt '- VaicitraYirya " : 30 men- 

2 Vide refs. ante. 

3 All of whom he went to the length of appropriating, apparently 

because Vali had originally wanted him to ^"beget sons on the 
chief queen Sudesnl and not on any other wife. 

4 Vide refs. ante. 

5 He was the favourite disciple ; Astavakra, and Svetaketu were of toe 

same age. and were "brought up as * brothers.' 

6 The case of Veda's wife seems to show that- * niyogas * were allowed 

during long abwnr?. of the husband, amongst rsi families at 

176 y 

apparent; so that In some cases priest influence (or a poly- 
andric tendency) can have been the only motive. The only 
clear case 1 of ' niyoga ' of widows is that of Vicitravirya's 
wives, another very probable case being that of Bhadra- 

Some curious points are raised by the case of Madrl. 
There it is said that inviting twins to a ' niyoga ' is tantamount 
to undergoing one only 1 ; Madri thus shrewdly secured to her- 
self the advantage of connexions with two persons and bearing 
two sons, while keeping to the letter of her co-wife's stinted 
permission! Kunti did not know this point of law, 2 otherwise 
she too would have, as she declared, invited twins every time. 
Such permission from the elder co-wife is not required in the 
case of Vicitravirya's wives, apparently as the 'niyogas' there 
were mainly on the initiative of Satyaviti ; but the Ko&ila 
princesses are also said to have been themselves very anxious 
for having children ; while in the other case, Pandu also him- 
self wanted to have ksetraja sons. The explanation is to be 
found in the nature of the permission Madri had. The 
notorious * mantra of Durvasas ' which Kunti allowed Madri 
to use once, suggests a regular system of * priestly bans ' or 
permits for various cases of royal license or transgressions. 3 
As already noted, Kuntfs ' boon ' or permit was not a rare 
one : at least two other well-known princesses of the Paurava 
race 4 having got it before her. Durvasas is said to have 
granted his absolving permit for free alliances (for progeny) 
4 in view of the coming age of distress/ referring obviously 
to the degeneracy of the dwindling Kanrava line, where sons 
proper could not be hoped for. 5 The " mantra ' probably 
consisted of some Atharva-vedic incantations 6 relating to 
amfcurs and fertility, 7 bearing the name and sanction of that 
noted Atreya priest, which forwarded privately to the selected 
quarters, produced the -desired effect easily, 8 by virtue of 
the authoritative license. 9 This ' mantra ' of "free-choice, 

1 It is to be noted that these details 3 and the statement that Sahadeva 

was born one year after Nakula, make it impossible that they 
^ were twins, unless that one year detail is wrong. 

2 This ^ probably also indicates a contemporary custom of twins some- 

times having a common wife. 

3 Instances of cheap absolutions of various sorts are not unknown in 

e tradition * ; e.g. from brahmanicide, parricide, adultery, etc. 

4 SatyaVati-Vasavl and Drsadvati-Madhavi-Yayatya , a similar permit 

seems to have been granted to meet the difficulties of Draupadi's 
5 successive marriages. Vide ante. 

5 As illustrated by the last generation, 

6 " Abhicarabhisamyuktam.": Mbh. I, 4386; 4748. 

7 Examples of which are well known in the Samhitas. 

8 N\rae, it is said, could resist the pharm. of the words of this ' mantra .' 

9 If licenses permitting; or ratifying questionable and irregular con- 

nexions could be granted to princes by Popes, in modern ages of 
criticism and unbeliefj they could very well ^iave been issued 
by as powerful and corrupt a priesthood, in a remote ancient 
age of credulous faith. 

( 177 ) 

therefore, being a special Atreya permit issued on behalf of 
Kuntl could only be lawfully used by Madri with' her own con- 
sent; thus she had to stop short, when that consent was 
withheld (even after Pandu's pleadings). and apparently 
Madri did not entertain the idea of a ' niyoga ' otherwise than 
by such romantic ' free-choice.' 1 

Of the f probable * cases, some are evidently so, only they 
have not been so designated in tradition, 2 while others com- 
prise a^ fairly numerous group, the chief feature of which is the 
ascription of the birth of sons unto old and childless kings to 
the propitiation and favour of some rsi who grants boons to 
their wives, and who often is the hereditary priest, 3 or 
continues to take an interest in those sons. This latter class 
as a whole, may or may not imply actual cases of * niyoga/ 
but a few of them undoubtedly were, in view of what has 
already been noted. 

Taking cases of the former group in order of sequence, 
we have, at the 22nd step from Mann, that of_ Purukutsa's 4 
queen obtaining a son for the race in the absence of her 
husband, according to the Vedic evidence 5 ; and this was 
apparently through a brother-in-law (who was also 
another husband, by biandry), according to the Puranic 
evidence 6 ; here 'niyoga ' and polyandry are combined. It is 
to be noted that this is also a Manva case, and the location is 
apparently the N armada region. 7 

Then, at the 44th step, is the famous case of ' dliarma- 
samkramana ' of Bh'aradvaja 8 into the Paurava dynasty. That 

1 Thus it would appear that In these cases we have a special type of 

e niyoga,' where the wife has free choice of any number of agents. 

2 As will be seen presently, this ohscurity has resulted from subsequent 

purposive handling of the traditional material. 

3 Cases of connections between such chaplains and the queens ^are 

numerous in Buddhist stories that refer to Puranic tradition, 
Vats. Ka. Sut. recognizes it fully in some courts. 

4 For the uncertainty of the identity of the Vedic and Puranic 

Purukutsas, vide ante, 

5 Vide ante. 

6 Vide ante. , . , , ,, .,* 

7 It 13 well known in ' tradition ' that the Manvas had spread thither 

aome time before the Ailaa, who subsequently intermixed with 
or absorbed them (cf. Pargiter : AIHT. p 256 ft). Mandhatr 
and Purukutsa's line seems to have thus branched off towards 
the S.W., among the Yadavas (like other Aiksvakas later on). 
Tliere were actually several Kosalas in the E., &, and b.VV., 
besides the N. and the Central, all colonized by Manva families 
of the same stock branching out at different stages in the dynas- 
tic sequence. Simply because a king is given in .the Aikwaka 
lists he need not be located at Ayodhya.-ascnpiion of Tule at 
that cite beinc a meaningless commonplace of later ages. 
8 Hariv 32- 172631; Brahml 13, 58-60; V*. 99.. 133-53; Mat. 
49, 11-34; Vi?. IV, 19, 4-8. 

( 178 ) 

phrase seems to refer to a e niyoga ' rather than to an 
' adoption >l for which the description ' samkramana ' would 
have been enough ; even that is hardly appropriate, and rather 
a rare way of describing it. The phrase yields better sense, 
with reference to the context and connected tradition, if taken 
to mean ' introduction of fresh blood (tainting, grafting) , 
through Bharadvaja, in accordance with sanctioned and right- 
ful custom (dharma) ', which explanation is added to many 
a traditional case of ' niyoga ' to justify it. It has already 
been noted that the contest and connected traditions above 
referred to, show that it was not Bharadvaja himself who 
became the successor of Bharata, but the son Begotten by him, 
Vidathin or Vidatha, or, as the Epic tradition has it, 2 
Bhumanyu, the son of Bharata' s queen Sunanda-Sarvasem- 
Kaseyl, born after the death of the nine sons. When it is 
remembered that the Kasis also were at that time under the 
influence of Bharadvaja-Angirasas, 3 as were the contem- 
porary Eastern Anavas of the adjacent regions, in which 
dynasty the Angirasa rsi blood hail 2 or 3 steps before been 
fully introduced (i.e., samkramita ' ) , and that the 
Sngirasas had come into close relations with the Paurava 
dynasty from the time of Marutta's adoption of Dusyanta, and 
also that the Angirasas had intermarried with Marutta's 
family, 4 the natural vagueness in the Bharata tradition 
disappears, and it becomes clear that the successor of Bharata 
s his ' ksetraja ' son 5 by his Angirasa priest. 6 

The next case is that of Vyusita^va, 7 whose place cannot 
be clearly defined. Ho is called * Paurava-vam^a-vardhana ' 
in the Epic story, but there is no Paurava Vyusita&va in the-; 
lists. The Aiksvaka Vyusitasva (at about the 81st step) is 
near enough to Pandu to be referred to as a precedent"; Kunti 
refers to Vyusitasva's story in reply to Pandu's citation of 

1 CI Pargiter: MET. P. 159 ff. 

2 Mbb. I, 94. 3710 ff. ; 95, 3785. 

3 Pargiter: AfflT. pp. 164, 220. 

4 Marutta's dtr. Samyata was given to Bharadvaja's uncle Samvarta, 

and another uncle (or step-father) married Bhad-ra, d'fcr. of 
' Soma,* prob. referring to an Aila or ' Soma ' king (who may 
be the above Marutta Taurvasa). Possibly both ' Samyata ' and 
1 bhadra xupena parama mata ' of the texts refer to the poly- 
androns Maanata (or * Mamata '). (The Tatirvasa and the Manva 
Marnttas were contemporaries, and both the neighbouring princes 
may well have been under the same Angirasa priest domination.) 

5 In that case Bharata's successors could well be called ' dvamusy- 


6 So also, the parallel between the brahmanical laudation ojf Yali 

Bharata is stTikine, 

/ mi. i, 121. 

( 179 }, 

Saudasa's case; even if (as is more probable) 1 this story were 
originally part of Pandu's exhortation illustrating * niyogas 3 
and wrongly attributed to Eunti, the point remains that it 
comes next to Saudasa's story, also an Aiksva-ka case; and the 
Aiksvakas were at this time well known to the Eoirus 2 ; thus 
' Paurava-vamsa-vardhana ' may be taken as a vocative 
referring to Pandu, and not to Vyusitasva. The story is 
apparently intended to show the superfluity of ' niyoga,' which 
it does not. The indications are plain that Bhadra-Xaksivati 
(an Angirasa lady descended from Kaksivant and Dirghatamas) 
had children after the death of her husband, and these were 
regarded in later ages as Ms children by a legal fiction, a- 
fact for which ' niyoga ' is quite sufficient explanation ; here it 
is no case of posthumous birth, for seven sons are born. 3 

The curious statement that the Paura-va (or Aiksvaka) king's 
sons were * 3 Salvas ' and ' 4 Madras/ is highly suspicious : 
it is obviously wrong ; the true reading would seem to have 
meant ' 3 from Salva, and four from Madia * respectively, 
that is of these 7 ' ksetraja ' sons 3 were begotten by 
the Salva and 4 by the Madra prince, who may have 
been relatives of the dead king chosen by the queen quite 
in accordance with custom. And all that is said about 
Bhadra's rising up from her husband's dead body and awaiting 
fruitful connexions in her own bedchamber, is strikingly 
similar in purport and details to the Bgvedic funeral mantras 4 
that make over the widow immediately to a relative of the 
husband. The queen's own descent being expressly traced 
from a famous ' niyoga, ' 5 resort to the same practice is quite 
intelligible in her case. 

AB has been noted above, some brahmanicised traditions 
assert that brahmans raised offspring on ksatriya widows 
after the Haih'aya-Bhargava conflict.. 6 This appears to be 
nothing but a polite way of saying what must have been 
natural, that the women of the beaten Haihaya chieftains 
were appropriated wholesale by their victors, and bore) them 

1 It seems that as Later on the propriety of Kuntfs making some pro- 

test was felt, the less known story of Vyusitasva was amended 
to form a reply. 

2 E.g. through Hiranyanabha and TJgrayudha, Santami's first wife 

' BhagTrathi/ the two wives of Vicitravirya ( f Kausalyas '), etc, 

3 And for 'that reason also s no case of begetting in sick-bed before 

death; the account also involves another absurdity of keeping a 
dead body nnburnt or nnburied^ for 7 years. 

4 Also to the case of Vicitravirya's widows. _ % 

5 Even if the Kaksivanb referred to is the Pajriya one ? the Angirasa 

connection remains, specially as the Pajriya Kakivant married 
the dauc&ters of an Mgirasa lady, Bomasa (w. of Svanaya). 

6 Of Mbh, I, 64, 2459-64; 104 } 4176-8; XIV, 29 ? 835 j etc.; cf. 

Brahmantja : III. 46, 30 ff. 

( 180 ) 

children; these victors were rather the Kasis and Aiksvakas 1 
than the Bhrgus alone; so that these fighting priests 
unscrupulously used the victories gained by their allies in this 
manner. At any rate this tradition does not prove prevalence 
of ' niyoga ' as a practice amongst the Haihayas (but 
indicates that the Bhrgu brahman s 2 also were conversant with 
the system, like the "Angirasas, etc.). For the Haihayas it 
only shows a taint in blood 3 owing to disastrous defeat. But 
it may well have been that powerful brahmans henceforward 
had an eye over the harems of chiefs: the instance of 
Dirghatamas-Angirasa having the use of a king's harem 
occurred just after this period of Haihaya-Bharga/va conflict 

and alleged wholesale ' niyogas. 

Iravat, 4 the son of the Naga princess Ulupi and heir to the 
Kauravya-Naga kingdom, was probably begotten by Arjuna 
by way of a ' niyoga.' One version makes her the widowed 
and childless ' snnsa ' of the Kauravya king, another the 
widowed and childless ' suta ' ; the latter is not probable, as 
Iravat is later on said to have been expelled 5 from the Kaura-vya 
court (where he was brought up by his mother) by his wicked 
1 pitrvya/ through hatred of Arjuna, though Iravat had been 
recognised as the heir to the Naga, kingdom. Clearly ' snta ' 
is an emendation in favour of Arjuna (the ' brahmacarin J ) ; 
there could be no case of a * putrikaputra ' succeeding (as 
with Citrangada) when a son was present ; it was his 
1 snusa ' whom Kanravya bestowed upon Arjuna for an heir, 
when she was herself desirous of offspring ; and the ( pitrvya * 
referred to was thus a younger brother of the deceased Naga 
prince, who was displeased at the prospect of a ksetraja son of 
the elder brother getting the throne. 6 Ulupi apparently was 
eagerly looking for a suitable ' agent ' with the permission of 

3- I.e. under the famous Pratardana .and Sagara ; probably also aided 
by Astaka (Kanyakubja), Usmara (Upper Doab and E. Punjab), 
and Vasumanas (Ko&la). See Pargiter : AIHT. pp. 268-71, 

2 Later on some cases occur of sons to kings being born through 

favour of Bhargava rsis (vide infra). 

3 But ace. to another version (Mbh, XII, 49) brahmans also were 

similarly tainted in blood during the Haihnya disturbances, 
while ruling families got intermixed with indigenous tribes like 
Rksas and Golan gulas. 

4 Mbh. 585. b. (Bhisma-vadha. ) : VI, 90; cf. 83, 3661. 

5 After "which Iravat repaired to Arjuna } then preparing for the great 

battle in the Himalayan regions, and joined in his enterprise. 

6 The parallel with the Dharfcarastra dislike of the Pandavas is 

striking! the want of this feature in the previous generation 
strengthens the presumption that Bahlika and Bhisma were 
the real progenitors of the elder Vaicitraviryas. 

( 181 ) 

her father-in-law, 1 when she met Arjuna at Gafigadvara, and 

persuaded him to accompany her to the Naga palace close bj 
and stay with her for a short time till she conceived of him. 2 
The details of her adventure and advances show that the 
initiative in the matter was almost- wholly hers, bearing a 
striking similarity with Saradandayam's case in regard to 
the quest and random selection. 3 

To the second group of the ' probable' cases belong the 
following in order of sequence 4 : The birth of Visvanutra- 
Visvaratha (Aila-Kausika) (31) through the favour of Eeika* 
(Bhargava) ; -of the sons of Sagara (AiksvakaJ (41) through 
that of ' Aurva ' (Bhargava) 6 ; of Damayanti-Vaidarbhl 
and her three brothers (Yadava) (50 or somewhat earlier) 
through that of a rsi Damana 7 ; of the sons of Ajamidha 
(Paurava-Bharata) (52) through that of ' Bharadva-ja ' 
(Angirasa-Bharata) 8 ; of the son of Dasaratha-Lomapada 
(E. Anava) (64) through that _ of Esyasrnga-Taibhandaki 
(Iiasyapa) 9 ; of Haryanga (E. Anava) (67 or 73) through 
that of Punarbhadra-Vaibhandaki (Easyapa) 10 ; of Jara- 
sandha-Yarhadratha- (Paurava-Magadha) (92) through that of 
Candakausika (Angirasa-G-autama-j 11 ; of the son of ' Srnjaya '- 
Pancala (bet. 66 and 93) through that of ' Narada ' (Xasyapa- 
Pancala) 12 ; of Draupadi and Dhrstadyumna (Pancala) " (94) 

1 It is remarkable that the Jataka tradition also attributes the birth of 

the ancient king Sagara-Brahmadatta of Kasi to such a quest; 
for a suitable consort on the part of a widowed and rather for- 
ward Naga princess, who met the exiled Kasi heir-apparent ; 
(vide infra). This is interesting for Naga ethnology. Uliipl is 
repeatedly asserted to be a dtr. of a Kauravya family ; so either 
she solicited a consanguinous * niyoga ' or the ^Tagas of Gangad- 
vara were Kurus as much as the Hastinapura family. There are 
many other illuminating statements about Nagas in the Epics 
and Puranas. 

2 Mbh. 248 (Arjunavana. ) : I, 214. 

3 Vide ante; it is to be noted that Kunti also was descended from 

Aryaka the (Kauravya) Naga (through female line). 

4 But not in order of probability. 

5 Va. 91, 64-89; Bralimanda. Ill, 66, 35-60; Br. 10, 29-50; Hariv. 

27, 1432-52; Vis. IV., 1, 8-15; Mbh. Ill, 115 (Jamad. ): 
638. b. (Eamop. ). ; ' 

6 Mat. 12, 39-42; so also Vis. IV, 3, and Brahmanda. Ill, 63 j cl 

Sagara and wives soliciting Vas'istha and then other rsis fcr 
sons : Br. 78, 3-11 ; cf. Mbh. 3H/ 106 ( s penances ' of * ionless 
Sagara & wives and ' boons ' to them). 

7 Mbh. 344 (Nalop. ) : HI. 53, 2077-80. 

8 VS. 99, 163-4; Mat. 49, 45-6. 

9 Va. 99, 104; Br. 13, 40-43; Hariv. 31, 1696-8; Mat. 48, 95-6. 

Mat. 48, 93-99; Hariv. 31, 1700-1701: vide n. 9 above; cf, the 

earlier case of Anga's birth in Bali-Jnava'a line. 
11 Mbh, II, 17 
n Mbh. VII. , 55; XII, 31. 


C 182 ); 

through that of Yaja and Upayaja (Kasyapa) 1 ; and of the 

children of Satadhanvan-Hardikya (Yadava) (94) through that 

of ( Cyavana ' (Bhargava) . 2 Apart from the old age of the 

' father ' and his childlessness, and the concerned rsis beltog- 

ing to clans otherwise associated with the * niyoga ' practice, 

the special circumstances that may have a bearing on the 

probability in each case, are : In the first : the alleged 

complete Bhargava priest-domination over Visvamitra's grand- 

parents, Kusika and his queen, 3 the visit of Paurukutsi 4 to 

the Bhargava retreat where she too 'Conceived along with her 

daughter Satyavati, 5 the continued Bhargava-Kausika 

' sambandha/ 6 and other previous and subsequent Bhargava 

intermixtures with royal families 7 ; in the second: previous 

Bhargava connection with Sagara's parents, the son- 

less and aged Vahu and his Yadavi queen, 8 and the alleged 

contemporary prevalence of ' niyogas ' of ladies of the ruling 

nobility lo Bhargava priests 9 ; in the fourth: the well-known 

.previous ' dharma-samkramana, ' of Angirasa-Bharadvaja blood 

into the Paurava-Bharata dynasty, 10 and the continued con- 

nection of the Bharatas with the so-called Bharadvaj,as, practi- 

cally a branch of that dynasty 11 , in the fifth (and the sixth) : 

the unique nature of the enticement of Ksyasrnga into the royal 

harem, 12 where his position is very similar to that of Dirgha- 

tamas in the harem of Lomapada'siancestor, and the continued 

Anava-Elsyapa eolmection shown by two ' boons ' of offspring 

1 MbL I, 167 (Drau. Sambh. ). But taken together with the Jataka 
version of Krsna's origin, this account would seem to be rather 

irl x A r , mal affiliation 'to Drupada and his sister-queen 
of the twin children of the Kos'ala queen abducted by him during 
pregnancy. At any rate Yaja is stated to have summoned the 
queen Parsati to receive offspring from him, and when she 
pleaded personal unpreparedness for the process, to have assured 
ner that her person would not be recuired, but that the offspring 
was ready for affiliation to her. The details thus show that 
-T T I ^ firs 4 *. x P ecte ! a ' ni yg a '> but the ritual arranged 
'"? a \3?* an l Bru P ada was found to be intended for 
chlldrei1 of the Kos'ala queen. 

of ' ni ys a ' arose wh ^ ^ ta - 


I TWTA" I u l ^- avano !?; U) : XJLJLI > 52 - 56 > S P- 52 and 55. 

5 Vis TV 7 n'i *? P ' A : -S 1 ' 49 ' ^ ff ' ; vide n - 5 below - 

*'lTT 'ffi'-^Kn * T\ku T 5 T T ; ^ ariv - 27 ^ 1432-52; Brahmanda. 
on r ^ " 01 '^ ^p - J JM-bn. XII, 115; vide n. 4 above. 
e n. above; ci. Br. 10. 63. 64-66- Hariv 27 14^7 
Va. 91, 97 ff., etc. ' J Mi) '' 

J -.8- "^ 011 |h Bevayani or Jamadagni with Aila or Manva 

139- Bralmlika Br "TIT ^'^* ^^^ IZ > 76 ff - 5 VS.?" 88^^120- 

Q , T ., V5>lln ' s w .soliciting Bhargava '). ' a ma ' If * l> 
9 Vide n, 6 3 p. 179. 

10 Vide ante, ^ pp. 177-79. f 

H Vide Patgit^r : AIHT. BB 112- 247^ 

12 Mbh in, 110, 9989 to ig, Mxfej Ram . I; 9 and 10 . cf 

iOj O. * * 

'( 183 X 

within a few generations 1 ; in the seventh ; the precedent of 
Vali and Dirghata-mas, where the rsi family concerned as well 
as the locality are the same, 2 and the notorious connection of 
the Grautamas with Girivraja from Vali's time to 
Jarasandha's 3 ; in the eigth'j similarity with the Visvamitra 
and Lomapada cases a in all three a rsi son-in-law 4 being the 
source of the son. 

Dhrtarastra's sons (94) are said to have been born 
through the favour of Krsna-Dvaipayana, on whom Gandhari 
attended to his satisfaction 5 ; but as it is also said that the 
* boon ' was a divine one, and elsewhere that the sons were 
directly begotten, 6 it is not a likely case of the above group. 
Ugrayudha (Dvimidha-Pancala) (90) is said to have belonged 
to the Solar dynasty 71 ',; his father or ancestor Krta (89 or 84) 
was a famous disciple of Hiranyanabha-Eausalya (83) 8 ; and 
that strange particular may be the indication of some infusion 
of esteemed Kosala blood into the Pancala family in the 
time of Krta. 9 

There is another fairly numerous group of cases, where 
sons are said to have been born to kings in their extreme old 
age, not through rsi favour, but by virtue of austerities and 
divine boons. 10 The instances of ' Asvapati ' of Madra (later 
than 30) , n Dirghatapas of Kasi (15) ^ Usinara of the Punjab 

1 Dasarathi-Caturanga and Haryanga were In the 7th and 10th steps 

from An^a. 

2 Vide n. 11, p. 181 Canda-Kausika here seems to mean son of a Kusa 

(a desc. of Kaksivat Gautama), who was called the ' Canda*; for 
locality cf, p. 165 ante, and n. 6 there.' 

, Vide n. 6, p. 165 and n. 11 p. 181, Cf. Mbh. II, 760, 802, 807, 886: 
III, 8083 ; etc. ; also XII, 168-73. 

q Narada was apparently also a cousin, of a collateral Pancala family; 
he is very frequently called a desc. of Paramesthin (vide Sor. In- 
dex, pp. 538-9), and P. "was one of the famed ancestors of the 
Pancala group of families (counting many rsi families among 
them : cf. Paramesthya rsis, Sor. Index, p, 539). N.'s connection 
with the Srnjayas and Pancalas, and his m. with S. s s dtr. (a 
cousin) is thus quite intelligible (as also this probable ' niyoga 5 ). 
6 6 Mbh. 180 (Samhh. ) : I, 115. Cf. I, 95, 3809; 110, 4371 ff. (Siva's 

' boon); 4378; 4522; etc.; but cf. I, 4558 (' arsah sambhavaji ' 

of the Dhartarastras). 

7 Mat. 49 5 61; vide n. 8 below. 

a Va. 61, 43-44; Brahmanda, HI, 35, 38-49; Vis. XH, 6, 4-7; also 
Hariv. ; vide n. 7 above. 

9 Note that the famous Yyusitasva * niyoga * occurred only 1 step 

above Hiranyanablia (83). 
It is possible that in some of these cases, the whole process and result 

10 oj: ' niyoga * are thus summarised and concisely put in an, 
acceptable manner. 

Mbh III, 293, 297, 299; his children were called by the metronymic 

11 Malavah ' : 297, 16807. 

Br 11, 36^7; Hariv. 29, 1522; Va. 92, 7-19; Brahmasfla, mi, 67, 

12 8.20, 

( 184 ) 

(26) / Pratipa of Kuruksetra (87) , 2 or Jyamagha of Vidarbha 
(38)', 3 are same of the best known In tradition. 

Connected with ' niyoga ' are some other cases of 'ksetraja' * 
sons (accepted as such )* known to tradition, technically belong- 
ing to the categories of ' gudhaja-/ ' kanma/ etc. It is note- 
worthy that the definite examples of such sons occur amongst 
ruling families chiefly in the period just before the Bharata 
war. 4 The queen of Ugrasena (Yadava) (90) , a Yaidarbhl- 
princess 5 while on a short ^isit to her father's capital and dis- 
porting herself in -his pleasure gardens, was beguiled by one 
Gobhila (who had impersonated Ugrasena) into cohabitation 
with him, and after the discovery of the fraud was abandoned 
by him in pregnancy; she then returned to Ugrasena, and the 
son she bore was Karnsa, ' son ' and successor of Ugrasena. 6 
Kunti's ' kanina J son, Kama (92), is recognized as ' by law 
the son of Pandu ' ; Krsna asserts it, and Kama himself 
acknowledges it, while Draupadi also agrees to this status ; 
in fact, on the eve of the great battle and after it Ivarna- is 
almost generally taken to be virtually the eldest * ksetraja ' 
son of Pancju and thus heir to the throne. 7 But it is rather 
different with Satyavati's ' kanina ' son (90), who is definitely 
* Parasarya ' 'and ' Parasara's dayadah, 3 and not Santanava, 
though he is claimed as elder brother of Vicitravirya through 
his mother and younger brother also of Bhisma, and is 
permitted the privileges that would have belonged to 
Santanu's own son. 8 Unlike Kunti's ' kanina ' son, her 
sister Srutadeva's apparently similarly born son Ekalavya 
(92) is not recognized as the Karusa king's own son; he too, 
like Karna, seems to have been abandoned, to be brought up 

1 Br. 13, 20-24; Hariv. 31, 1674-9; Brahmancla, III, 74, 17-20; etc, 

2 Mbb. I, 97; v. 3882 seems, to state that {Santanu was so called being 

the son of p. after his death, i.e., by 'niyoga'; v. 3799 is in 
explanation of the frequent statement that S. was Maliabhisa- 
Aiksvakava, * reborn '; for ' mahabhisa '=the great healer; this 
again seems to conceal a real dynastic fact that P. 'B w. bore 
to M. by ' niyoga ' a sou San'tanu-Mahabhisa, who also married 
a * Bhagirathi * or Ikavaku princess, previously promised to P. 

_ or married to him. 

* Tis. W, 12, 2-15; Hariv. 37, 1981-9; etc. 

This shows that in court circles ' ksetraja ' sons by lawful ' niyoga ' 
had become so common that those by occasional illegitimate 
agencies did not rouse much comment ,and could become quietly 
affiliated. The view that the Panqlu ( niyogas ' were a Himalayan 
novelty, becomes therefore untenable. 

5 Prob. Bhismaka's sister; having slain the son of a Vaidarbhl princess, 

Krsna naturally incurred the enmity of Bhismaka and his 

6 Padma IL 48-51. 

7 Mbh. V, 137; 140-43 (sp. w. 4734 and 4756); cf. 144-46; also VI, 

43; 122; XII, 27B, 817 (cf. 427/ 3 1488; XI, 27), Karna is called 
' K-uru-vira, 5 ' Kuru-mukliya/ etc., along witr^ Arjuna. also by 
himself j e.g., VIII, 4925. 

8 Mbh. VI, 594; cf. XII, 350, 13643; XIII, 18, 1341; vide pp. 

168, 169, ante. 

( 185 ) 

by a Nisada chieftain, but not subsequently affiliated like 
him to the mother's legal husband. 1 According to the 
Buddhist version of Epic tradition, 2 Draupadi and her brother, 
instead of being the miraculously born children of Drupada 
and PrsatI, were born of the Kosala queen, who was carried 
off in her pregnancy by the victorious king of ' Easi s (i.e. S. 
Pancala, with which it was at that time often amalgamated, 
as also with Kurus) after her husband's fall in battle, and 
gave birth to those children (twins evidently) as wife of the 
latter king, who accepted them as his children. Apart from 
greater credibility, this version remarkably confirms the 
contemporary view of ' ksetraja ' -children, as exemplified in 
the case of Ugrasena and others. The story of Matanga, 3 
son of a ' brahman! J by a low-class paramour, who was yet 
regarded as the c son ' of the ' brahman/ is not definitely 
ascribable to any particular time. 4 According to the 
' B'amayanic ' tradition, ' Hanumant ' (65) was such a 
' ksetraja ' son of Kesarin, recognized as son in spite of 
illegitimacy, as were also the 2 sons (Vail and Sugriva) of 
Rksa's wife. 5 

1 Vide p. 158 ante* and n. 1 there. It is possible that another sister 

of Kunti, Srulakirtti, m. to a Kekaya king, also underwent 
some sort of polyandrous ' niyoga ' : she too had 5 sons who were 
expelled from sovereignty by other Kekaya princes and joined 
the Pandavas to regain their kingdom (cf. Mbh. V. 22, 664 and 
Sor. Index, s.v. Kaikeya, etic. \ also Va. 96, 145 if; Vis. IV, 4, 
11; etc.). 

2 Cowell : Jat. V. 225 ft'. ; vide n. 1, p. 182, and n. 2, p. 132. 

3 Mbh. XIII. 27-29. 

4 But the mention of Matanga' s austerities and influence at Gaya and 

the very name of Matanga, probably indicate some close connec- 
tion or identity with the famous Trisanku-Matauga (32), the 
' CarKjala, 1 of Kikata, whose father Tryaruna was a Bgvedic 
rsi. Cf. Nanda parallels later on. 

5 Mbh. Ill, 147, 11193 ff. ; also Bam.; Bralimanda, III, 7, 212-16; etc. 

The traditional evidence regarding widow-remarriage 
as we'll as widow-burning, is comparatively meagre ; no very 
early instances can be discovered ; it is only towards the close 
of the period covered by the traditional sources that several 
definite cases, crop up. Evidently, during the greater part of 
that period, widow-remarriage was more or less taken for 
granted, or .was superfluous owing to prevalence of ^niyoga/ 
or customary transference, to elder or younger brothers-in- 
law, and no special social conditions had arisen 1 to develop 
3, custom of widow-burning. It is striking that most of the 
early instances of this latter practice refer to the Manvas and 
in a. less degree to connected Yadavas ; while the later 
instances belong chiefly to the Yadavas, and partly to one or 
two other closely connected families : it seems possible that 
with the early Manvas the practice was original, being a relic 
of not uncommon primitive beliefs and institutions, while 
the later Yadavas adopted it under the stress of the struggles 
and disasters of their age. It is also remarkable that 
remarriages of widows or ' equivalents ' are almost exclusively 
noticed in their prevention or subsequent non-occurrence, 
except occasionally where the woman concerned is either a 
* naga,' 2 dasyu,' 3 ' vanara ' or c raksasa,' 4 or primarily in the 
possession of some ' asura- ' (etc.), who is killed 5 ; which 
would indicate that those few preventions and these latter 
mixed cases were the exceptions to and special cases of a 
general and therefore unemphasised usage of remarriage. 
The distribution of the references also points to such 
remarriages being commoner in the Madhyade^a (Kuru- 
Pancala and Eas'i-Kosala) towards the close of our period. 


In a number of instances the line between widow- 
remarriage and ' niyoga/ or ' brother-in-law marriage/ or 
even polyandry, can hardly be clearly drawn,: e.g., in the 
already noticed cases of TJlupi (94), Ambika and Ambalika 
(92), Bhadra-Kaksivati (81? or 42?), Purukutsam-Narmada 
(22), Tara and Mandodari (65). But if they are not unalloyed 

1 Like continued foreign invasions and 'domination in the medieval 


2 E.g in the case of TJltipi (already noted), and that of the Kas'i king 

Sagara-Brahmadatta's mother, who was a ' Naga ' widow re- 
married. (Jatakas : Cow^ll : VI, 81). 

3 E.g. in the case of the Gautama who married a * dasyu ' widow in 

an Eastern country (vide infra). 

! St ^ cases of T5irS and Mandodari, noticed ififra and elsewhere. 
6 Vide i 

( 187 ). 

widow-remarriages, they clearly imply that custom for those 
times. In another group of cases there is no proper ' re- 
marriage of a widow,' but rather ' re-connection of an 
equivalent of a widow ' ; e.g., where on ousted, or vanquished 
chieftain's wife is approached or appropriated by the victor, 1 
or where a queen sets up her paramour as the king after 
murdering her husband, 1 or where a dowager queen goes to 
live with a chaplain, 1 or where fair ladies are abducted by 
' assuras * etc., and rescued, either forthwith or after long 
stay with the abductor, by heroes who subsequently marry 
them. Gases like these, as well as the readiness with which 
the claims of several princesses to restored maidenhood were 
admitted, and their easy and normal subsequent ' remarriages/ 
show that, of the later objections to widow-marriages, a 
principal one had little force in those days. 

Taking the few probable indications of and direct 
references to such ' re-connections ' and re-marriages all 
together, they are found to be thus distributed regionally and 
by groups : 

Amongst the Manvas : Bhalandana/s eon Vatsapri (8), 
of the Vaisali line, rescued the youthful Mudavati, daughter 
of his father's friend King c Viduratha ' of the Nirvindhya 
region (evidently an early Yadava), from her abductor 
Kujrmbha, whom he slew; he then married her, though she 
had lived with that Kujrmbha for a considerable length of 
time. 2 Several other similar instances are known to Vaisaleya 
tradition. 3 

1 Such cases are frequent in the Buddhist versions of the Puranic tradi- 

tion. Of. Coweil : Jatakas : VI, 244 (a 1ST. Pancala ^case) V, 225 ff. 
(Brahmadatta's widow living . with his chaplain ; also a N. 
Pancala case) ; in the first instance there is a fnll remarriage, 
the murdered king's son calling the paramour (step-)f ather ; the 
first and third varieties are however indicated by Epic-Puranic 
statements themselves (e.g. re N-ahusa and Ugrayadha, and re 
Brahmadatta's wife. 

2 Mark 113 ff. 

3 In Mark.Piir. 113-36. Thus Aviksit (39) married the Vidisa 

princess Vaisalini-Bharnini after rescuing her from an abductor, 

whom he slew (he had declined to marry her before, having been 

defeated before her at her sv'ayamvara, where he had seized 

her) i again, Dama, (42) married Sumana-Basarni after she had 

been seized from him by the Madra and Vidarbha princes, whom 

he slew or defeated and thns rescued her. The Mark. Pur. 

also gives (31-35) a Kasi parallel to Vatsapri's case^ where 

Pratardana's son and Alarka's father Vatsa-Etadhvaja (42^ 

rescued Madalasa from the inner apartments of the abode of 

her abductor, and married her after a romantic adventure. o 

an uncertain but a remote early period (pre-Manva, (referring to 

Auttami-Manu) the same Pur. ascribes (69-72) two instances 

of abduction, of TOtama's q^&en and of a ' brahmani/ ana tue 

subsequent smooth restoration of both to their husbands. 

( 188 ) 

The famous Aiksvaka, Satyavrata-Trisanku (32), appro- 
priated to himself the newly-married wife 1 of a \idarbha 
prince (Yadava), 2 whom he apparently slew 3 m battle with 
his supporters, and had by her a son VisnuYiddha. It is said 
that the capture was not quite illegal (though disapproved by 
Kino- Trayyaruna and ' "Vasistha '), as the marriage was not 
yet "technically complete; still the point remains that he was 
regarded by many to have virtually made another's wedded 
wife forcibly his own, and also that he was by some others 
thought to "have been unjustly banished for such capture, 
which was actually common in dynastic history even in later 
times 4 ; the legal point of the ' 7th step ' 5 is apparently a later 
gloss to justify the great Yisvamitra's support of Trisanku : 
for the completion of the early Vedic marriage did not depend 
on the 7th "step/ 5 but on ' pani-grabha ' of the bride and on 
subsequent home-coming and consummation. Trisanku also 
seems to have similarly appropriated an ordinary citizen's 
wife, 7 unless the two notices refer to the same facts. 

Etuparna-Aiksvaka (51, or 42?) is connected inseparably 
with the ancient and genuine Nala-Da-mayanti tradition 
(Yadava) ; according to this, 8 Damayanti, in order to find out 
whether the reported new charioteer of the Kosaln. king was 
Nala himself, despatched messengers to Etuparna's court to 
inform him that she had decided to hold a second ' svayam- 
vara ' very shortly, no trace of her missing husband being yet 
found; and Iltuparna at once set out for Kundina to have his 

1 Brahmancb : III, 63, 77-114; Va. 88, 78 fi. 

2 An anachronism : apparently a slip for Vidoha (Manva), quite a 

common error. Qr Vidarblia may have been used by anticipa- 
tion here, and means only * Yadava.' 

3 Prob. the unmeaning ' hatva divaukasam ' of the texts stands for 

4 hatva Vidarbhakam ' or ' Videhakam. > 

4 Cf. Ugrayudha preparing .to take away Santanu's widow; and tbe 

several references in the Jatakas to ancient kings of Kas'i or 
Pancala similarly abducting the queens of other defeated and 
slain kings. 

5 All the Purlnas do not agree -with regard to this ' 7th step ' 5 some 

have * paiji-grahana mantras * instead. 

6 Besides, as the marriage which Trisanku interfered in was evidently 

a ksatriya one 5> it must have been complete (cf . Mark. Fur. " 
13-36 3 where in conn, with Dama's marriage with Sumana, 
these points are discussed in the Svayamvara assembly) by the 
simple step of stating mutual consent, "or placing the wreath 
of choice, or grasping of hand in defiance of assembled 
ksatriyas, -before Trisanku carried the bride off. Even after 
such completion, diverse ' ceremonies ' of marriage were gone 
through, as in the Epic svayamvara&, but that had only a social, 
and no legal value; 'these ceremonies might be performed long 
after consummation. Prob. it was during such secondary 
ceremonials that Trisanku abducted the bride, and thus plainly 
violated the marriage., and took to wife an _actual * punarbhu ' 
or * anya'mrva.' 

? Brahma : YT !, 98 ff. ; Hari'v, 12, 717 fL 

8 Mbb. Ill, 62^79, sp. 69-77. 

( 189 ) 

chance of obtaining Damayanti as wife. Damayant! adopts 
this ruse with her mother's consent (though her father knew 
nothing about it), and it leads to no graver consequences than 
a pretty little scene of lovers' pique, and Etuparana's polite 
apology. It is quite evident from this case, that in both the 
Vidarbha-Yadava and the Kosala-Manva circles remarriage of 
widows or ' equivalents ' was not discountenanced, and were 
quite ordinary occurrences. 

For the time of Eama-Dasarathi (65) there are two mixed 
cases of widow re-marriage (as already noticed) amongst the 
aboriginal (but civilized) races of S. E. Deccan, connected 
with the Manvas (viz., of T'ara and Mandoda-ri, with Eama's 
approval); and one reference to a possibility of a similar 
' mixed ' case amongst the Manvas themselves, as between 
Sita-Vaidehi (or-Kaus v alya) and Latsmana (or Bharata) , in 
the event of Eama's death. 1 

With this group may be placed the case of Gautama 2 

(Angirasa), who married -a ' dasyu ' widow (bestowed on him 

by a ' dasyu J chief), settled amongst her people, and had 

many sinful children by her. The personal and topographical 

details in the story 3 show that it is evidently another version 

a sarcastic and a Western ,anti- Angirasa one of the famous 

tradition of Dirghatamas' adventures in the eastern countries. 

This ' dasyu ' widow may or may not be the same as the 

Sudra ' Usij or Ausinari of the better-known versions ; in 

fact Dirghatamas took to wife a number of such women from 

Vali's harem or capital, by whom also he had numerous 

children. If identical, the * dasyu ' woman of the former 

version need not be taken as a ' widow,' but rather as f one 

separated from (or kept separate by) her husband,' for 

* bhartra virahita ' can mean both ; and the latter meaning 

would suit the case of the transferred Ausinari quite well. In 

any case the Gautama-Angirasa Dirghatamas (41) had a wife 

who was either a ' full ' widow or a clear ' equivalent/ or had 

two wives of each description. 

1 Vide ante. 

2 Mbh. 658b (Krtaghnop. ) : XH, 168-73. 

3 E.g. the 'dasyu' chief = the Anava Vali (often confounded with the 

Daitya Vali) ruling over non-Alias; 'the great Baksasa city of 
MeriWraja = Girivraja (once the city -of the Eak. Bsafcha, ace. 
to Mbh.); the (< neighbouring ' Baka ' king on the Ganges 1 ' 
agrees with the topography and Epic tradn. ; the benefactions of 
the patron goddess Surabhi of MeruVraja agrees with the story 
of Snrabhi's grace on the prolific Gautama at G-'invraja ; the 
sinful sons of Gautama are the disdained e Krsnanga (or 
Kusmanda) Gautamas of Girivraja in the usual versions; tins, 
and the "award of hell to Gautama, show, that the version arose 
with a croup hostile to Angirasa pretcnsipns^and Narada (fcfee 
narrator)^ being apparently a Parame?thina-Panca,la, may well 
have represented such a group and tradition. 

( 190 ) 

Amongst the early Alias, the case of Pururavas (3) who 

rescued the .-.bducted Urvasi, and marrie'd her, 1 and the 

statement in the brahmanicised tale of ' Indra-vijaya ' that 

Nahusa courted the queen of ' Indra ! after expelling and 

succeeding him, and was on the point of having her as his 

wife, 2 are the only indications of an acquaintance with 

* punarbhii '-or ( anyapiirva '-marriages in that age. Later 

on, however, Ajamidha (53) seems to have had a * punarbhii ' 

wife, Dhumini, from whom the main Paurava line was 

descended. 3 

Amongst the later Ailas, Tadavas and Pauravas, more 
definite ~and frequent references are found. Some of the 
earlier Yadava cases are also Manva cases, a noted above ; the 
clearest of them being that of the proposed re-marriage of 
Damayanti-Yaidarbhi (51, or 42?). The next indications are 
much later, belonging to the period just before the Bharata 
battle. The ' 16,000 wives '* of Ersna (94) belonged 
originally to Naraka-Bhauma of Pragjyotisa, having beon his 
entire select harem, which Krsna captured and transferred 8 
to Dvaravati for himself, after slaying Naraka. 'As already 
noted, such harem transfers were not rare in this or subsequent 
ages, _ and it necessarily involved * equivalents ' of widow- 
marriages. Thus Vatsyayana's Sutra clearly records tradi** 
honal dynastic customs when it classifies 6 the king's 
' antahpurani,' in order of court precedence, as the ' clevis, 1 
being his own normally wedded and principal wives the 
punarthw, 9 evidently the widows of the predecessor and of 
vanquished kings, and the courtezans, both ' abhyantarika ' 
(i.e., those taken into the ranks of the ' zenana ' wives) as 
well as 'natakiya' (i.e., the court actresses and 'dancing 
girls) 7 Arjuna's (94') settlement of Vajra (97) and Satyaki 
and Krtavarman'a (grand- or grea.t-grand-) sons, with the 

^ Pu ^ r A a , vas ' ,/eputed grandfather Soma's appro- 
sriet Wlfe T and his nine other similar 

: 90, 28 

Matsya : 50. 

2 Mbh. 

3r*$ jt 


4 The figures of course are not to be taken literally. The Jataka 

tradition knows of other ancient kings with ' 16,000 ' wivea (vide 
infra sec. re polygamy). v 

5 Not improbable as Krsna may have easily penetrated into N. E. 

Rajgal after Magadha had been subjugated by Jarasandha's 
In later history cf. the parallel of Gurjara kings carrying off 

i.l TlYn KPO. M Q Cf ' -P-HT-itTrt f^^-iiJ^ /Oil- _ 1 /M 't nr^ , ** O 

a Vats. 5 r 5 T 4 aUa ^ ^ ^ Cent A '^ 

7 Separate private chambers and gathering halls are assigned in that 

whert V^t 6 T? 1V Q S -f f ^ ^ K "r* be not ^ d that else ' 
where Vats Ka. Sut. recon ' ^ 

V TQ- r that ' 

Vats Ka. Sut. recognizes ' punarBhii^ wives as normal 

and frequent, apart from court circles. (e.g, IV, 2 31-44) 

( 191 

remnants of the Yadava seraglios, 1 must have involved similar 
4 punarbhu ' marriages. What is indicated for group in the 
above cases, is illustrated for individuate, in the story of 
Pradyumna and Mayavati, the widow of Sambara. 2 

Amongst the later Pauravas all such references belong to 
the period immediately before the Bharata battle, and are 
comparatively .clearer : Karta-Ugrayudha (Dvimidha-Pancala) 
(90/91) wanted the widowed Satyavatl-Vasavi (Pauravi) to be 
his wife, 3 within a few days of Santanu's death, and sent a 
messenger to Bliisma demanding his step-mother; as Bhisma 
did not agree to such an ill-timed and ill-worded demand, 
Ugrayudha invaded Hastinapura to enforce it, but was killed 
in battle. It is to be noted that the main objection of Bhisma 
was that the proposal was haughty and inconsiderate, the 
funeral ceremonies of Santanu being yet unperformed ; nothing 
is said regarding the inadmissibility of the widow-re- 
marriage involved; and if Ugrayudha had won the battle, he 
would have married SatyavatI quite in accordance with the 
royal custom 4 of having ' punarbhu ' wives. 5 Satyavatfs 
marriage with Santanu, was also an equivalent e anyapurva ' 
one, she having borne a son to ' Parafora * (Vasistha) before 
that; and though the fact may not have been known to 

1 Mbh, XVI, 7, 230-'53. 

a Br. 200201; Hariv. 163168. 

Other indications may be found in the cases of : (1) Rukmlni, who 
being first betrothed to Sis'upala was almost an * anyapurva ' ; 
and Sistrpala too would gladly have taken -her as wife If he 
could, after her marriage with Krsna. (2) Bhanumati, who was 
violated by Nikumbha of Satpura, and lived, with him. for a 
long time, and was after her rescue married to Sahadva- 
Pandava. (3) Bhadra-Vaisali, who may either have been a 
' shared ' wife of Vasudeva, the Karusa king, and Sis'upala, or 
a widow of one of the latter two, finally taken into the seraglio 
of Vftsud-eva. 

3? Hariv. 20, 1085 1112: she was also sought in marriage by Asita 
after her amour with Parasara, apparently while still living 
with him : Mbh. I, 100, 4045. 

4 Of. the ' punarbhu * wife of Ugrayudha's ancestor Ajamidha. 

b That this practice was known to the Pancalas is also proved^ by the 
' Jataka ' statement that Draupadi's mother was the widow of 
the Kosala king, after whose defeat and slaughter she was 
during her pregnancy married by Draupadi's putative and step- 
father, and made his chief queen; other similar cases are known 
to the Jataka tradition, referring to S. Pancala (taken as= 
Kas'i) and Kogala, and to the Epic and post-Epic periods. In 
Paficala, again, Mahaculanf s queen married her brahman, paramour 
after her husband's murder, which is parallelled by what Kama 
(in the Epic) says about the minister Mahakarna appropriating 
the quen of the expelled or slain Magadba King Ambuvlca. 
In a later period, nearer to Buddhism, a woman of KoaZa 
prefers a brother's life to a husband's, as other hushands and 
children Jby them might be obtained afterwards (Cowell : 
Jatakas : I, 165). 

( 192 ) 

Santanu, it was quietly accepted later on by the Kuru court; 
but it was unlikely that the brahman Asita-Devala (Kasyapa) 
did not know of it when he stood a candidate for Satyavati's 
hand. 1 Very similar is the case of Kunti-Sami (Yadavl) 
(92/93J, whose previous ' connection ' also was apparently 
unknown to Pandu, but subsequently an open secret in all 
court circle. Satyavati herself had no scruples regarding the 
permissibility of widow-remarriages : immediately after 
Vicitravirya's death, she proposed that Bhisma (90/91) should 
marry the widows of her on, the Kosala ( = Kasi) princesses; 
Bhisma's non-compliance is amply explained by his famous 
pact, whereby- he could not many and have children who 
would inherit his claim ; it is to be noted that this did not 
legally prevent him from accepting an invitation to ' niyoga ' 
(which lie probably did), for in that case it would not be his 
rights to the throne that would be passed on to the begotten 
son, but that of his * putative ' father ; thus it becomes 
intelligible why Bhisma should have put forward the counter- 
proposal of ' niyoga ' of the widows as better than their frank 
remarriage. ^ While however Ainbika and Ambalika could well 
have remarried, their elder sister Amba is rejected by the 
Salva king (Marttikavata-Yadava) after her forcible abduction 
by Bhisma, in spite of her assertions of innocence ; it is said 
that Salva did not like the idea of marrying an ' anapfirva ' 2 ; 
but that is not enough of an explanation in view of contem- 
porary ^ dynastic practices; ' dread of Bhisma/ so often 
emphasised in the Amba story, together with some amount of 
vanity, was ^ plainly the greater part of the reason. As noted 
already, it ^is probable that, as amongst the Yadavas, there 
were occasional harem-transferences amongst the Pauravas 
also, 3 and it seems likely that some of the Dhartarastra widows 
(94/95) ^ were taken into the Pandava seraglios as the * punar- 
bhu s wives of the five joint kings. 4 

As noted above, the earlier references to widow-burnin^ 
belong to the Manva group : Thus the Vaisali king 
Khanitra's (20, or several steps later) three devoted wives are 

1 Mbh. I. 100, 4M5. 
a Mbh. V, 175, 5979 fl 

3 Of. the proposed^ transference of Vicitravirya's wives to Bhisma; or 
what Urvasi says about the ancient custom of Pum* princes 
approaching a predecessor's wife; the case of Pratlpa (89), 
aanga-Bhagirathi and fiantanu, may have been really a case of 
snch transference; cf._the parallel Yadava instance of Jyamagha 
(38), the captTired pnnoess, and Yidarbha, where also th^re is 
the same probabihty. Of. Cowell, Jat. VI, 133, for a clear case 
of harem transference from the Magadha King Arindama to his 

same as the contmp ' of " 



( 193 ) 

said to have died along with their husband, with whom they 
had retired to the forest in old age; how they died is made 
clear by the subsequent case of the retired Vaisali King 
Nansyanta (41) and his wife Indrasena who ascended the 
funeral pyre of her husband when he was murdered by a 
Yadava king, Vapusmat, in his forest retreat. Such 

sanamarana ' was not however fully customary in this 
family, for a few steps above, Vira, the queen of Earan- 
dhama (38), continued in her austerities for several years after 
her husband's death in their forest hermitage (within a 

* brahman ' settlement) . l In the Aiksvaka family 2 there is the 
well-known instance of Bahu's (39) Yadavi queen, who was 
on the point of committing ' suttee/ but was dissuaded by her 
Bhargava benefactor, on the ground oi her pregnancy and 
prospect of the birth of an auspicious son. 3 

f-4 /m gr Up also falls the case of --- 

]iti (62) , an Aiksvaka princes, married to Jamadagni-Bhargava 

(daughter s son of another Aiksvaka princess), who wanted to 
immolate herself on the funeral pyre of her husband after his 
slaughter by the Haihayas, but was prevented by the 
Bhargava elders. 5 It is noteworthy that no ' brahman ' case 
of 'suttee ; is known to tradition, and this is quite in keeping 
with B the difference between * brahman J and ' ksatriya ' social 
life in general. The case of Eenuka cannot be taken as a 
brahmanical ' one, as she is herself a ksatriya princess 
married into a mixed ' brahma-ksatra ' family with pro- 
nounced ^ksatriya traits. The only ' brahman ' instance 
probable is that of the Ingirasi 6 in the Ealmasapada story 
(54) ; but the episode seems to have been either wholly 
invented, or to have been subsequently used, to explain 
Madayantl's ' niyoga,' and naturally modelled on the Panda 
story ; so that it is the self-immolation of Madn that is echoed 
in what is said about the IngirasT. Even if this be a genuine 
case, the Manva connection is obvious. 

Long after these Manva and connecte'd cases, we have 
several others amongst the later Yadavas and Pauravas : 

On Pandu's (93) death, Eunti (Yadavi) wanted to be 
burnt with him, being the elder queen; but Madii dissuaded 

I These details are in Mark. Pur. 113-136. 

3 If the Vyuitasva of Kunti's story is the Aikvaka Vyusitasva, then 

the case of Bhadra-Kaksivati may be placed here, as one of 

prevented ' suttee ' for the sake of progeny, 

3 B<J. Ill, 63, 126-33; Va. S3, 120-39; Br. 8, 29-46: Via. W, & 

15-18; etc. 

4 Bd. Ill, 30, 34-50; etc. 

5 It Is rather striking; that in both these cases of dissuasion it is a 

Bhargava. "who dissuades. 
Mbh. I, 182 ? cf. VI?. IV, 4, 19-38 j etc. 

( 194 ), 

her, showing the fitness of her following the dead husband, 
and ascended the funeral pyre. But that is only one version 1 : 
the other and more probable version 2 (because it is interwoven 
with details of subsequent event's) shows that Madri simply 
died a tragic death soon -after Pandu, apparently from grief 
and shock, and the two unburnt bodies were (on the 17th day 
after Pandu's demise) brought down from the hills to the 
Kuru capital, where they were burnt together with royal 
honours and ceremonials : so that there was no case of widow- 
burning here, only a touching synchronous death. But the 
point remains that both the queens wanted to commit 
' suttee * actually, and there was no dissuasion from that 
purpose . 

The curious statement, that after the slaughter of 

* Eicaka ! his kinsmen obtained the permission of his brother- 
in-law Virata (Matsya-Paurava) (93) to force Draupadi (the 
Sairandhri) to mount the funeral pyre of ' Eicaka ' and be 
burnt with his body, 3 shows that amongst the Kicakas there 
was a practice of burning the favourite woman of a dead 
chieftain along with him. Who these Kicakas were is not 
very clear; they are not derived from any Aila or Manva 
family by tradition ; they may have been Kikatas ; at any rate 
they were closely connected with the Matsyas and Trigartas, 
and had intermarried with the former ; and the Matsya king, 
having sanctioned the proposal regarding an attendant of his 
own court, must have been fully aware of such a practice. 

It is remarkable that in all the detailed account of the 
general destruction of the Eurus (and other combatants) in 
the Bharata* war, there is not a single instance of ' suttee.' 4 
There are however several * equivalent J cases, those of 
suicide by drowning, in connection with the tragic episode of 

* Putradarsana ' ; a number of Kuru (and Pandava-Pancala) 
widows (95) plunged into the Ganges, with the permission of 
Vyasa, to follow and rejoin their dead husbands, of whom 
they had had a night-long spiritual vision through the grace 
of that sage. 5 Ulupi (94) later on followed apparently the 
same -course, when she * entered the waters of the Ganges/ 

i Mbh. I, 125. 

2 Mbh. I, 126-27. 
5 Mbh. IV, 23. 

4 Thus women go out of the city into the battlefield to bury their 

dead husbands and relatives : there is no ' sahamarana * : XL 11, 
298; cf. 1 and 9. 

5 Mbh. XY, 

( 195 ) 

on the ' mahaprasthana ' of her husband Arjuna, etc., 
tantamount to death, to the world, 1 But that phrase may also 
mean that (like another wife of Arjuna) she returned by river 
to her Naga father's riparian principality on the Upper 
Ganges. The other Pandava wives left behind did not think 
of any rash steps : Citrangada repaired to her son's kingdom, 
and Subhadra and the rest remained with Pariksit. 

The later Yadava cases are the clearest of all, strikingly 
similar to the medieval Bajput (Indo-Scythic) * Jauhars * of 
the same regions and the alleged ancient Scythic custom. 
This may have something to do with tEe early close 
connection between the Haihaya-Talajanghas (Yadavas) and 
the Sakas, 2 etc., jointly with whom they raided and spread all 
over N. India. The Sakas and kindred tribes were indeed 
already settling in (the submontane) part of the kingdoms of 
Kanyakubja and Ko^ala, having intimate political relations 
with the Kusika- Alias and the Vafigthas 3 ; and they seem to 
have settled in and dominated Kosala for a long time before 
Sagara (41) (for about 18 generations), forcing some of its 
ruling families to branch off in different directions. 4 It is 
remarkable that the Manva instances of widow-burning should 
be confined precisely to this period, and to those branches of 
the Manva stock that still clung to Kosala and the adjacent 
Vai^ali in spite of Saka (and Haihaya) domination. The 
obvious inference is that under pressure of the circumstances, 
or as- a result of prolonged Saka influence, the Manvas (of 

1 Mbh. XVII, 1. 

a Re the Sakas in Puranic tradn., vide end of sec. re sister-marriage. 

3 JKuSika-Kusasva (29) "lived amongst the Sakas, probably after 

Kanyakubja had been overthrown for the first time by \adava- 
Haihayas between Sasavindu and Durdama'a times (or by tne 
Sakas themselves), and it was with their support that he re- 
covered the throne; a few steps Later, his descendant Visvamitra 
Visvaratha (32) was discomfited by Vasistha's Saka allies 

4 It is noteworthy that in about the same time (22-34) as the Yadaya- 

Haihaya and Saka invasions (20-41), all the noted /ik*vaka 
groups of kings are located in regions _other than Ayodhya: 
e*r Purukutsa (22) migrates to Malwa m the S. W. (which 
suggests that the powerful Diruhyus, counting Sakas, etc. amoBgst 
thfm whom Mandhatr (21) had molested, now invaded Kosala 
inTetaUatTn) where his and his brothers' descendants continued 
m retaU S Haryasva, whose wife was the notorious 

ri seems to have migrated into the guraaena-Yadava country 

( 196 ) 

whom the Sakas were an early branch according to tradition) 
often adopted the (kindred) Scythic practice : thus it Is that 
DO further cases are found amongst them aft&r this period. 
On the other hand, the Sakas, etc., in the latter part of their 
predatory career became definitely associated with the 
Haihaya- Yadavas, and after Sagara's time apparently became 
merged in their numerous ranks, 1 being humiliated and 
expelled by that king. This would explain the few earlier 
Yadava-Manva cases of this period, and also the subsequent 
reappearance 2 of Scythic-like ' group ' ' suttees ' amongst the 
later Yadavas in an age of disasters. 

These ' group suttees ' occured, it is said, 36 years 3 after 
the Bharata battle, when the confederate Yadava clans of 
Dvaravati perished ingloriously in internecine strife; urged 
by Satya-SStrajiti, Krsna recklessty completed the destruction, 
and then met with his death apparently in the course of a 
Itfisada raid on the helpless city; old Vasudeva (93), died of 
grief; thereupon the 4 favourite wives of Vasudeva, Eohini- 
Pauravi (of Vahlika), 4 Bhadra-VaWali, Devaki and Madira, 
mounted his funeral pyre 5 ; Eukmini and some other wives of 
Krsna (94) also did the same, 6 but others like S&tya-Satrajiti, 
etc., retired to the Himalayan hermitage of Kalapagrama (as 
did the widows of Airura) ? It seems however that Eukmini, 
etc,, did not forthwith ascend the actual funeral pyre of 
Krsna but entered the fire several days afterwards, 8 probably 
when the Yadava cavalcade hurrying to Indraprastha was 
surprised by the Abhiras. 

1 Just as various other Manva sections were assimilated by the Yadavas 

from time to time. 

2 This would involve a supposition that the Yadavas preserved traces 

of Scythic admixture for about 600 years, which is not im- 
probable as that admixture itself took about 250 years; besides 
the Yadavas had subsequent connections also with the Drhuyus 
and other Western peoples, to whom the Sakas are traditionally 
traced; thus there was another aka-Yavana invasion of Mathura, 
from the Himalayan regions in the time of Krsna; and Krsna's 
exploits brought him in touch with the Western nations spe- 
cially (as the Mbh. says). 

3 Prob. only 6 years. 

* It is noteworthy that both Eohini and Madri are princesses of the 
Vahlika dynasty, a N. W. offshoot from the Kurus of Hastinapura, 
and both "had Yadava co-wives 3 whose example they apparently 

5 Mbh. XVI, 7. 

6 Ibid. Bukmim-Vaidarbhi, Jambavati-Kapindraputn (cf. XIII, 629), 

Saibya, Gandhari, and Haimavati, the last tbree hailing from 
the N. W * 

7 Ibid; cf. retirement of Kuru ladies; Mbh. XV ? 15-18. 
a Ibid. 

( 197 ) 

So far only the positive references to widow-burning have 
been dealt with; but it is to be noted thafc there are many 
more instances in traditional accounts, where details of the 
demise of kings (and brahmans) are given, but where either 
nothing is isaid about the 'sahamarana' of their widows, or they 
are plainly stated to have lived on normally : e.g., in the cases 
of Dasaratha and Vrhadvala amongst Aiksvakas, of Earan- 
dhama amongst Vaisaleyas, of Karna amongst, E. Anavas, 
of Bharata, Santann, Vicitravlrya, the Dliartarastras, 
Duhsala, Abhimanyu, etc., amongst Pauravas, of Ugrasena, 
Kamsa, Akrura, etc., amongst Yadava-s, or of Auddalaki- 
Sujata and Saktr's wife Adr^yanti, amongst Angirasas and 
Vasisthas. Thus no general custom of ' sahamarana ' is 
proved for any particular group, though in some of these 
groups more cases can be discovered than in others, owing to 
circumstances and facts explained above. 


It is supposed by some Vedic scholars that polygamy was 
dying out in the ' Bgvedic ' period, yielding place to mono- 
gamy. 1 The Bgvedic evidence by itself, however, can be 
made to prove either this or the opposite theory. The correct 
interpretation of that evidence, therefore, must depend on the 
historical data supplied by the traditional accounts. These 
make it plain that polygamy never died out in any part of the 
Vedlc age, but existed before it, during it, and after it ; only 
there were several well-marked periods and groups in which 
the practice was more in evidence than in the others, owing 
to circumstances not unintelligible; and at the same time 
there were other groups and periods more or less characterized 
by monogamistic ideals. On the whole, however, polygamy 
was associated with ' brahman ' influence, dynastic expansion, 
or frequent wars, and was on the increase towards the close 
of the Vedic age ; while monogamistic tendencies were more 
apparent amongst the non-brahmamc, or minor, or temporarily 
unimportant principalities. 

Quite naturally, a number of the cases of polygamy 
noticed in tradition are also those of concubinage, 'or posses- 
sion of slave-girls, or connection with courtesans, or general 
laxity in sexual morals.. About these instances also, the 
same remarks hold good, regarding their developing conditions 
and distribution in groups and periods. 

Polygamy is ascribed- to 'some of the kings of the pre- 
Aila and pre-Manva dynasties. 2 Whatever may be the value 
of these semi-legendary references, that of the uniform 
ascription of group ' polygamy to the * prajapatis ' for a few 
generations immediately before the rise of the Manva and 
Aila dynasties, 3 is clearer : these latter cases seem to belong 

1 Vide ante re Vedic evidence on this point. 

2 E.g. Uttanapada had two wives Sumti and Suruci, mothers of the 

famous Dhrava and Uttama respectively (' Svayambhnva ' 
vams"a, in almost all Puranas). The Mark. Pur. ascribes 4 
wives (1 chief and 2 companions -f 1 temporary) to ' Svarocisa ' 
(f. of Svarocis.a ' Manu), whose polygamy is denounced, so that 
he finally becomes an ascetic (Mark, Pur. 61-68). The same 
Pur, ascribes six royal wives to King Durgama of Priyavrata's line 
(f. of c Raivata * Mann) (ibid. 75) ; and Svarastra (f . of ' Tamasa, 
Manu) also is said to hav-e had other wives besides the chief queen 
(ibid. 74). The first two cases are located in the N. Himalayas; 
the 3rd in the S. W. (Anarta, Surastra, etc.), and the 4th on 
the Vipasa (N. W.) ; these were amongst the regions occupied by 
the Manvas and pre-Manvas before Aila expansion, 

3 In all Pur., with more or less details. It seems probable tliat 

' rajan/ * manu/ * prajapati ' or ' indra ' (z& known to Pur. 
tradn.), were different types of mlers, with distinguishable 
features and periods of flourishing. 

( 199 ) 

to the first stages of Manva and Aila race-settlements, when 
wholesale transferences of the daughters of the one clan 1 
(probably conquered) to the possession of tfie other horde- 
leaders (' prajapatis ') would be natural. 

In the very next stage , polygamy is found continued 
amongst the early Manvas, but only rarely and gradually 
amongst the early Alias. Manu (1) had ten wives according to 
brahmanic tradition 2 ; that of the Epic and Puranas does not 
mention it, except once, but ascribes to him as many, or many 
more sons/ Iksvafoi 4 (2) and Vikuksi 5 (3) had very large 
progenies, who were settled in several groups in Uttarapatha 
and Daksinapatha, and in the N. Himalayan and submontane 
regions, respectively : their polygamy is obvious. So is that 
of Vrhadasva (11) and Euvalasva (12), to each or one of 
whom from 100 to 21,000 ( !) sons are ascribed 6 ; and of 
Yauvanasvi-Mandhatr (21) , who had (besides four noted chil- 
dren) 50 daughters, 7 all of whom he bestowed on a Kanva (Pau- 
rava) rsi, Sobhari (who had 150 sons by them) , 8 just as three 
or four 'steps above, Mandhatr's mother's ancestor Eaudrasva- 
Paurava (17) had given away all his ten daugEters to another 
opulent rgi, the Atreya Prabhakara (from whom the 

1 E.g. the daughters of Daksa, of whom 13 were given to KaSjrapa, 

27 to Soma, etc. Instead of being taken as obscure cosxnogonic 
myths, this common Puranic account may be taken as stating a 
tradition that when the Prthuite dynasty of Daksa was supplanted 
by Manvas and Ailas, the women of the former royal family 
passed into the possession of the Manva and Aila leaders (typi- 
fied by ' Kasyapa * and ' Soma. J ) 

2 Mait. Sam, I, 5, 8; cf. Mbh. XII, 13596. 

3 At ithe commencement of all Solar dynastic lists; the number varies 

bet. 9 and 10, ace. as 'Ha' is included or not; one version 
gives 60 more sons to Manu, who perished in dissensions: 

4 2+50+49=100 sons : Vayu : 88, 8-11. 20-24; Brahma : 7^ 45-8, 61; 

etc.; and Mbh. XIII, 2, 88; also Matsya : 12, and Padma : V, 

5 114 (in (1 the &T+ ; 16% the N.)=120 sons : Matsya : 12, 26-8; Padma : 

u HI, 63, 29-62; cf Mbh. Ill 202 

beg?) and 2(J4 (beg.); also Brahma : 7, 60-86; Hariv. 11, 674-5 

Vi iBi 1 fc, 38-55; Patoa : VI, 232, 16. 

aa in several other cases. 


Svastyatreya families thus descended) - 1 It seems possible, how- 
ever, that the bestowal of those fifty maidens was regarded 
as a gift of sla^'e-girls by that rsi 2 ; this would then indicate 
that these ' daughters ' of Mandhatr were bom of captive 
or slave-girls acquired in the course of his wide conquests. 3 
A sister of a * Yuvangtsva ' (apparently some near descendant 
of Puruku'tsa who had Naga connections) 4 had five daughters 
by a Naga king Dhumravarna of the S. W. littoral (or 
adjacent islands), who were all given in marriage to the 
already married son of Haryasva-Aiksvaka (27), who was 
ousted from Ayodhya, and settled amongst the Yadavas of the 
S. W, 5 

On the other hand, amongst the early Ailas the first 
probable -case of polygamy is that of Nahusa (5), who (Besides 
his sister-wife Viraja) 6 is said to have coveted ineffectually 
the queen of an ' India ' whom he displaced, and otherwise 
assumed a sensual turn of mind after his rise to great power 6 ; 
but this does not prove a practice' of polygamy. With Yayati 
(6), however, it is polygamy distinctly, connected with con- 
cubinage and relations with slave-girls, 7 Yayati had no harem 
before his Bhargava marriage : for he was then single, and 
had subsequently to build special apartments in his palace 
for Sarmistha and her 1,000 or 2,000 ( !) companions and 
attendant slave-girls. He obtained Sarmistha and three other 
maidens by .virtue of his marriage with Devayam, whose 

1 Particulars in Brahma : 13, 6-14 ; Hariv. 31, 1661-8 ; Vayu : 99, 

121 ff. 

2 Ev. VIII, 19, 36, is usually so interpreted, though ' Vadhu ' there 

may as well have the sense of ' wedded * wives or brides (being 
carried home). 

3 As in the case of his father-in-law Sas'avindu: vide infra. 

4 Visrra : IV, 3, 6-12 ; this Yuvanas'va may be Ambarisa's descendant 

J23 ?). 

5 Harivamsa : 94-95, 5142-5206 (Vikadru's account of Yadava expan- 

sion, told to Krsna). 

6 Vide ante. 

7 Apart from his 2 chief queens (in all Puranas : e.g. Vayu : 93, 28 ff. ; 

Brahma: 12, 22 ff. 1 ; Hariv. 30, 1601 ff. ; Matsya : 25, 6; 27-32; 
Visnu : IV, 10, 1, etc.) and a host of concubines (cf . Mbh. I, 
80-82), tradition credits him with further amours in old age, 
' e.g. .with Visvaci (Visnu: IV, 10, 6-8; Brahmanda : III, 68-70), 

OT As'ruvindumati-Eatiputri (Padma : I, 76-81), for whose 
sake he wanted to kill those two queens and quarrelled 
with his sons (piob. for the second time). (Note. ' Ratiputri * 
'seems to be^a modification of ' Raji-putri ' which was not under- 
stood ; ' Eaji ' was the ' Indra * in Nahusa's time, and he may 
well have used his daughter in ruining his rival and brother's 
son of a neighbouring principality ; Nahusa's coveting ' Indra's ' 
queen thus would mean simply his overthrowing his brother 
Baji and toying^ to appropriate his wife, quite a ' real * event 
to be recorded in tradition and to be referred to by Salya by 
way of an example to Yudhisthira. ' Asruvindumati ' is not a 
likely proper name ; but * Vindumati is ; so the original phrase 
would seem to have been * agra Vindumaiti E'aji-putri ' or ' the 
unmarried daughter of Raji named Vindumati.') 

( 201 ) 

bondswomen they had become as a result of the Bhargava 
priestly domination over Vrsaparvan's family. Sarmistfaa 
established her status (and thereby that of the rest of her 
party also) as Yayati's. wife by explaining to him that a 
bride's female companions and slave-girls were also lawful 
wives by custom, being bound to her and being given away 
at the same time ; sh reminded him that she was thus given 
away by Vrsaparvan, who is elsewhere stated to have enter- 
tained Yayati at his court on the occasion of his marriage. 1 
From these details it would at first seem as if Yayati's poly- 
gamy was derived as much from the Bhargavas as from the 
4 Danavas ' ; but even before Yayati Ayu had married a 
' Danava ' princess without any indication of similar circum- 
stances 2 ; so also later on Puru, Dh'undhu and Dusyanta, all 
sons of ' Danava ' mothers, 3 were either not associated with 
polygamy, or where so, it is not directly attributable to the 
4 Danava ' source 4 ; no instance of polygamy is found amongst 
these earlier (and real) Danavas, whereas several ate recorded 
about the early Bhrgus (viz., the * Bhrgu ' who married 
Daksa's daughters, 5 Cyavana-, 6 and Sukra himself) 7 ; and in 
this particular instance the Danava kingdom and dynasty 
was completely dominated by the Bhrgu priest. It is also 
significant that it is with Yayati that ' brahman ' influence first 
effects a real entry 8 into the Aila group, after several prominent 
attempts and conflicts. 9 His son Puru (7) may possibly have 
had two wives 10 but no clear case of polygamy occurs again 
amidst the early Alias (Pauravas or Yadavas) till the time of 
KaudiaSva (17) u and Sasavindu (20), with both of whom 

1 For the above details, vide Mbh. I, 78-83, and the Pur. refs. above, 

2 In most Pur., e.g. Vayu : 92, 1; etc. 

3 This generalization is made by the Mbh. and the Pur. themselves in 

the genealogical portions. Dhundhu is called ' Danayusa's ' son ; 
evidently referring to a similar parentage as in the three other 
Aila cases. 

4 For Dusyanta's probable case, vide infra. 

5 In most Pur. accounts of Daksa's progeny. 

6 Arusi: Mbh. I, 66, 2605 ff. ; cf. Ill, 174, 65; Sukanya : Mbh. Ill, 

121, 10313/122, 10320-44; 124, 10371 f . ; IV, 21, 650-51; V, 116, 
3970; Va 86, 23; Bd. Ill, 61, 19; Pad. IV, 14, 49 f. ; Bam. 
V, 24, 11 ; numerous other c Kanis,' and Vadhus/ ace. to KV. 
and Bra. (for refs. vide ante,) r 

7 Go (a sister): in all Pur. accounts of Pitr-vamsa ; Jaynati (d. of 
7 either Nahus* or Eaji) : Va. 97, 149-54; 98, 20 j Bd. HI, 72, 

150-6 ; 73, 19 j Mat. 47, 114-21, 186. 

He gets his ' victorious car ' and wives from a Bhargava brahman, 
of whom he is afraid and who curses him, and then restores 
him to favour, even sanctions the unusual succession ot Furu, 
which is therefore accepted by the people. (His brother *ato 
became a 'muni* under Manva influence apparently). 

9 E.e. in the time of Pururavas and Nahusa. *!- 

10 One a Kau^alya : Mbh. 156 (Puruvams. ) : I, 95, 3764; another, 

' Pausti ' : ibid. 94, 3495. M,-^-, . 

11 As already noted, his 10 dtrs. were married to one ri V 3 *' 

and he had 10 sons also besides those 10 dtrs. : Brahma ; 
6-14; Hariv, 31, 1661 ff.j etc. 

( 202 ) 

' brahman ' and Manva co-nnect-ions are obvious. The 
victorious and conquering Sasavindu is credited with such a 
large number of sons and wives, 1 that many of these latter 
must have been simply concubines or war-captives and slave- 
girls; such full-fledged polygamy cannot have 1 cropped up 
suddenly, and if there was a previous history, it would be 
fully explained by the continued connection of the Bhrgus 
(associated with Yayati's polygamy) with Tadu's descend- 
ants 2 who also quite early became rich and aggressive, and 
developed predatory tendencies, 

In connection with this Haihaya-Yadava expansion^ 
Indeed, there is a marked frequency of polygamy in all the 
groups concerned with it : it seems as if with the Haihaya- 
Yadavas themselves, the polygamy was due to their raids, 
conquests and opulence, while with the rest it was partly 
due to the tribal needs of struggling against numerous hordes 
for about two and a half centuries, and maintaining dynastic 
strength and continuity, and subsequently, to the flush of final 
victory over the invaders. The period coveted by the Haihaya- 
.Yadava troubles and the coeval cases of polygamy is that 
between the outer limits of Sasavindu to Bharata (2044) 
and the inner limits of Krtavlrya to Sagara (30 41). 

The first Haihya-Yadava case of Sasavindu (20) has 
already been noted; his sons (21) also were evidently poly- 
gamous, being ascribed large progenies, 3 which in fact is a 
main characteristic of the whole race ; so were also Aijuna 4 
(31), Talajangha 4 (34) and Vitihotra 5 (36), with their 
hundreds of sons ; the patron-priest of the Kartaviryas (31) , 
Datta the Itreya, was likewise a pfolygamist 5 ; while Supra- 
tlka (39) and Durjya's (40) several wives 7 and their sons are 

1 Brahma: 7, 94 (Vindumatfs 10,000 younger brothers); of. 

Brahman<f a : III, 63, 70-71 ; Hariv. 12, 712-13 ; Vayu : 95, 20 (100- 
hundred sons); but Matsya 44, 18 . (100 sons); Vinu : 
IV, 12, 1-2 (1 lakh wives and 10 lakh sons) ; cf. Mbh. 595 
(o<Jasa-rajika) : VII, 65, 2321 ff. (the ' sacrificing ' king, had 
100,000 wives, each mother of 1,000 sons). 

2 Yadu was ' cursed ' to live amongst his mother's people and follow 

them, Jt'.e. the Bhrgus). There was also some maritime 
Naga influence, far * Yadu ' or an early Yadu prince married 
Slaughters of such a Naga Sea-king over and above his other 
wife or wives, and these Nagas, had inter-married with Manvas : 
Hariv. 94, 5193 ff. 

E.g. Vayu : 95, 20; prob. this is referred to in ' Sasavindavl prajah ' 

which filled the earth. 
* Hariv. 34, 1890-1, and 1894; Vayu: 94, 46-49 and 51, etc. 

He had as numerous a progeny as Talajangha, the two groups being 
mentioned often together (e.g. Brahmangla : III, 47, 68 ff.); 
"besides he had some descendants * as a Bhlrgava brahman * subse- 
quently. Ace. to Mbh*. he had 10 wives and 10ft sons, a moderate 

6 Mark. 17 and 18; Padma : II, 103. 

7 Karh having two queens at the same time : Varaha ; 10, 17-34 and 51-67. 

< 203 ) 

named in story. 1 It thus seems probable that other famous 

Haihaya leaders, like Krtavirya, Durdama, Bhadrasrenya, 

etc., were also polygamous. This general Yadava tendency 

is illustrated in another fresh offshoot of the race in the same 

period: Jyamagha 2 (38) had evidently a strong inclination 

towards polygamy, in spite of his precarious life of poverty 

and struggle, and once captured a maiden in a victorious 

raid, 3 but could not marry her for fear of his queen, a Saivya 

(who had devotedly shared Ms exile) ; his son ' Vidarbha ' 

(39/40) inherited this captive princess (probably not the 

only one), and had at least another wife 4 who bore him 


The cases in the other group belong to the Aiksvakas, 

Vaisaleya-s, Kanyakubjas, Kasis (the restored) , Pauravas and 

E, Anavas, with connected Angirasas/ most of whom were 

affected by the Haihaya movement. Amongst the Aiksvakas 

of this time Trisanku (32) had at least 2 or 3 distinct wives 

[viz. , Sa'tyaratha of Kekaya 5 (mother of Hari^candra) , the 

captured wife of s Vidarbha ' 6 (mother of Visnuvrddha) , and 

of ( a citizen s ] 7 , and his 4 harem ' 8 is said to have been under 

* Vafigtha's ' control during his exile. 9 Later on Bahu (39) 

has also at least two wives, hostile to one another 10 ,; so again, 

his son Sagara (40/41), who may have had many more wives 

1 It is to be noted that the particular naming in these two cases seems 
to be due to close connection with the Kasis, within the sphere 
of regular bardic chronicles; the pore Yadava genealogies were 
often deficient in particulars, and ii> is only for the periods when 
the Yadavas were brought into intimate contact with North- 
Eastern life -that d-etails regarding their genealogies become 

2 This famous story is in all the major Puranas; Vayu: 95; 

Brahman^a : III, 70; Matsya, 44; Brahma: 15; Visriu : IV, 12; 

3 This points to* one of the main sources of Yadava polygamy, captives 

and slaves. . 

a Bhima, etc,, were apparently her eons: Hariv. 37, 19e. 
5 Vayu : 88, 116-17; Tkhmanija ; III, 63, 115-16; Hariv. : 13, 754; 

Brahma : 8. *,-,*,* 

fi Vayu : 88, 78-79 ; Brahmanda : III, 63, 77 if . 

I Brahma : 7, 98 it ; Hariv. 12, 717-21. 

n Brahma: 8, 1-23; Hariv. 13, 728. , . T . , v * 

I The fact that Trisanku was ousted from his kingdom after his cap- 
ture of a Yadava ('Vidarbha') princess (cf. similar occasion 
for Haihaya-Vaisaleya conflict), and that he made friends with 
the expelled Kanyakubjan Visvamitra m exile, and jointly with 
him struggled to some sort of power again, shows that , theae 
prices were combating the Haihayas who had allied with the 
akas and Vas'isthas, and were over-rmimng the country from 

10 Brahmand: 111, <A 126-133; Brahma: 8, 29-46; and in corr. 
tions of Vayn, Visnu, etc. 

( 204 ) 

than two, 1 if there is any basis for Ms ' 60,000 ' sons 2 ; this 
is not improbable, as the chief wives of these two princes were 
taken from (branch) Yadava families, 3 where concubinage 
was not rare ; besides the very circumstances of struggle with 
the Haihayas and victory over them would induce similar 
polygamy by reaction, and by this time the Bhargava 
brahmans too had become associated with these Aiksvaka 

f * 

princes. 4 

The first (and apparently the last) Kanyakubjan (Aila) 
case of polygamy occurs with Visvamitra (32), referred to 
above, who had a number of sons by several wives already, 5 
when he begat Astaka on a Paurava princess (Drsadvatl or 
Madhavi) 6 to succeed him in the apparently temporarily 
restored principality of Kanyakubja. 7 So a-lso the only 
probable cases of polygamy in the Kasi (Aila) line are those 
of the struggling Divodasa-Bhaimarathi 8 (32 ?) and the 
victorious Alarka (43) , 9 in the earlier and the latter parts 
of the same Haihaya period. The only West Anava instance 
known also fall within this period: Uslnara 10 (bet. 26 and 
32), the first of this Western branch, had four royal wives 
and sons by them before he begat Sivi on. the same Paurava 

1 Matsya: 12,39-42; Vismi : IV, 3, 1-3 - Brahma: 8, 63-72; 

Brahmandla: III, 63, 154-9; Hariv. 13, 760-14, 807, etc.; the 
names ^ of the 2 wives are variously given, but they were from 
2 sections of the Yadava groups of the S. W. 

2 Cf . refs. in n. 1 ; e.g. Brahma : 78, 3-11, etc. 

3 Vide refs. in n. 1 and n. 2. above. 

4 Vide same refs,; both Bahu and Sagara were supported and be- 

friended by Bhrgus in their struggles; the Bhrgu? had become 
connected earlier with the Aiksvakas immediately after their 
expulsion by Haihayas, in the time of Jamadagni (32). Three 
generations after Sagara, Bhagi'ratha is said to have given away I 
million damsels out of his stock to brahmans. He was under 
Angirasa influence, Kautsa taking his dtr. to wife : Mbh. VI L 
2249 ff. ; XIII, 6270. 

5 Cf. the Visvamitra genealogy: e.g., the sons Hiranyaksa, and Renu 

by Salavati, Galava and Mudgala by another wife; etc. ; Brahma 
10, 5567; Hariv. 27, 1460-62, etc.; one wife of V. was 
supported by Tris'anku. 

% Mbh. Galavacarita ; Brahma: 13, 91-92; Hariv. 27, 1473; Brah- 
manda: III, 66, 75; Vayu : 91, 99103. 

7 Visvamitra had been expelled from it by Vasisthas and Sakas, 

prob. joined with Haihayas; two generations later the kingdom 
finally succumbed. 

8 T,he probability is suggested by details in Galavacarita, Mbh. 

9 So the details of Alarka's career in Mark. Pur. (25 E.) would 

suggest (cf. the common Puranic statement about * the youne 
and beautiful Alarka,' etc.). 
10 The details of his family are given in : Brahma :J.3, 20-24- Hariv 

S\ 1674 "To ; JSrir. 99 ' 17fi: Brahmandla ; XII, 74, 17-20 j 
Matsya ; 48, 15-21 ; Visnu : IV, 18, 1 ff. ' 

( 205 ) 

princess mentioned above. 1 The emigration of the ^Anavas 
from the Upper Doab in two divisions towards Punjab and 
Bengal may well have been due to the Haihaya (and Saka) 
attacks of this time; and as In the c^se of the other minor 
affected dynasties, here too, Ufinara's polygamy would be an 
indirect result of the impact. The Eastern branch, getting 
settled beyond Vaisall, apparently escaped * Haihaya ' influ- 
ences, but soon adopted the polygamy and lax harem-life of 
the Angirasas and pre-existing Manva-Saudyumna people (as 
already noted) ; thus the notorious cases of Vali 2 and the 
connected Angirasas 3 also fall within the period in question, 
though forming a separate group owing to difference in develop- 
ing conditions. The Vaisaleya (Manva) and Paurava 
(' Eestoration ') cases, however, are associated with both the 
sets of circumstances indicated above : on the one hand, the 
Haihaya invasions considerably affected them, the Vaisaleyas 
finally tiding over them after a struggle, and the Pauravas 
being almost crushed out at first but recovering after a long 
time; on the other, Vaisali was particularly an Angirasa 
sphere, and the restored Pauravas were also completely 
brought under brahman influence, first Kasyapa (Kanva) and 
finally ingirasa (Gautama-Bharadvaja) 4 ; thus in these cases, 
viz., of the Vaisaleyas, Khamtra (20 or later) r ; Aviksit (39) 
and'Marutta (40), and the Pauravas, Dusyanta (43) _ and 
Bharata (44), the effects of long and successful war with a 
polygamous race, as well as of Manva-brahman traditions and 

1 The historical basis of the remarkable story of D?sadvatI-Madhavi 

seems to be an emergency entente between the Paurava, Kosaia, 
Kasii, Kanyakubja and W. Anaya princes in the face of a serious 
common danger from the Haihayasy-whereby their dynastic 
continuities and solidarity of alliance were assured by -mam 
of Madhavfs special polyandry (cf the case of ^upadi). 

2 Vali evidently had a large * harem,' out of which 2 _wives aie 

separTey mentioned; viz., Sudesna and Anslnari; another 
distinct wife is probably indicated in the reference to a 


XIX, e.g. 

( 206 ) 

influence, have to be recognized. Of these princes Aviksit 
(39) 1 was not content with the six wives ;WQ became his 
by self -choice, and developed a princely hobby of carrying 
off princesses holding their ' Svayamvara ' and thus filling 
his harem (a peculiarly Yadava trait) ; and his capture of a 
Yadava princess led to a concerted attack (apparently a 
Haihaya invasion), which was resisted successfully by his 
father Karandhama. Marutta followed his father in having 
a number of royal wives 2 ; besides he was a particularly rich 
prince, who rose to 'samrat'-hood by wealth alone, while 
others had to fight for the rank.. 3 Dusyanta's polygamy is 
evident from such statements as that Sakuntala consented 
to union with him only on the condition that her son should 
become the heir to the throne by preference, and that a 
number of women gathered to ssee him off as he set out for 
the hunt. 4 Bharata's three queens are well known in tradition. 5 

Notices of polygamy become very rare 6 in hg period 
that follows. It is possible that it was now somewhat 
discouraged, after the Aila modification of Manva brahman- 
ism/ begun by the Kausikas (32) and furthered by the 
Bharatas (44) ; but the already noticed cases of polygamy 
amongst these two groups would rather show that; in this 
respect they failed to change- the earlier traditions much, at 
least at first. A more probable explanation of this rarity of 
instances is that this period, unlike the preceding one, was 
not marked by any great prolonged wans and invasions, 8 or 
tribal and social upheavals, and was a comparatively quiet 
one, during which the several dynasties sank into petty local 
existences. 9 The few known cases of polygamy for this interval 
of about 5 centuries (45 85) must therefore be regarded as 
sporadic recrudescences, sometimes associated with philo- 
brahmanic^ princes and their priests. Prom the negative 
point of view, it is -also to be noted that this was a period 

1 The details of Vaisali manriage-relationa are in Mark, Pur. 113 

136, and generally agree with well-established synchronisms, 

2 7 wives and 18 sons ; names and parentage of these princesses are 

given in all the cases, 

3 Mbh. II, 15 (where claims of previous great kings to ' samrat J -&hip 

are compared with Yudh.'s), 

4 Mbh. (gakuntalop, ) : I, 73 j 69. 

5 In all Pur. genealogies. 

6 Only^ about 8 cases in 40 generations, compared with about 36 cases 

in the previous 44 generations. 

7 Vide Pargiter: AIHT. Chap. XXVI. 

8 The Bgvedic battles of this period were local, fought mainly by 

the petty Pancalas with neighbouring princes, with no great 
general results ; what loomed large in the vision of the Efcgvedic 
brahmans, was but an ordinary epoch from the wider' stand- 
point of Ksatriya tradition as a whole. ** 

9 Cf . the isolation of ihe Kos*alas, Pancala-Knrus, and Yadavas, for the 

greater part of this age (e.g. in the Ramayanic period). 

( 207 ) 


within which most of the monogamistic episodes of tradition 
fall, 1 and also one wherein some of the most important and 
properly 2 Kgvedic personages flourished. 3 Of the instancea 
referred to, three or four only can be assigned to definite 
steps in the dynastic scheme : Dasaratha (64) had at least 
three wives well known in- tradition, two of whom 
(' Eausalya ' and Sumitra) were ladies of his own family 4 
and court (being respectively a Hosala princess and a daughter 
of the Vasistha (or Angixasa) priest Yamadeva by a * Karana * 
wife), 5 and the third a Eekaya (W. Anava) princess, who, 
coming from a count with other traditions, 6 created troubles 
by seeking to appropriate the Mug to herself and ignore his 
other wives and their children; Datfaratha enjoyed the 
patronage of Yamadeva, a e Vasistha ,* and the Kasyapa 
Ttsyarnga of Anga (Ms son-in-law), to whose favour was 
due the birth of Ms sons (as ia his friend's case), 7 
Dasaratha's friend of Anga, and his namesake, distinguished 
from him as Lomapada (64) (East Anava), was also evidently 
the lord of a motley harem 8 ; the story of his childlessness and ' 
invitation of the Kasyapa brahman Esyasrnga into his harem 
for the sake of offspring is almost a replica of ffche episode of 
Vali and Dirghatamas; it is clear that the same degenerate 
harem life and conditions continued in the E. Anava group 
under Manva-brahman influence, only Lomapada improved 
upon it by associating a large number of courtesans? intimately 
with the court 10 and even the princesses, 11 thefirst 'deffnit* 
instance of royal recognition and employment of prostitution. 12 
Ajamidha (53) (restored Eaurava) had at least three wives, 
of whom one, DMmim was a * punurbhii/ which probably 
implies that he had a ' punarbhu ' section in his harem ; the 
Paurava court at this time was dominated by the same 
Angirasa-Bharadvaja priesthood as in the days of Bhirata (44) , 

1 All connected with rather isolated and minor dynasties.? and parti- 

cularly with Vidarbha and West-Anava principalities. 

2 Other c Rgvedic ' stages may be called the Manva, Kausika, and 

Early" Bharata (or Angirasa) ones, and this stage, the Later 
Bharata or Pancala. 

3 E.g. of the Sndasa group. 

4 Vide ante. 

* 5 * Snmitra Vamadevasya babhuva karam-suta. 

6 For there was a monogamistic tendency amongst the W. Anavas. 

7 Earn. I, 11. For Lomapada : Vayu : 99, 104, etc, 

8 Mbh. ill, 110-113 (cf. XII, 234, 8609 and XIII, 137, 6269); 

Ram. I, 9 and 10; Visnu : IV, 18, 3. Cf. also n. 7 above. 

9 Some of whom were hereditary. 

10 The ministers are conversant with them. 

11 Santa engages them at the command of <fcfae king; and Esyasrnga 

and his courtesans, are all kept within the inner apartments of 
the palace. . 

12 Other probable instances are much earlier : vide infra. 

18 Nilim, Kesini, and Bhumini : Brahma: 13, 81-82; Matsya : 49. 
44 ; Hariv. 32, 1756 ff . and 20, 1055 ; etc. 5 in some accounts 
the number is four, 

( 208 ) 

as shown by Ajamidha's obtaining sons through the 
favour of a Bharadvaja (clearly a cousin-priest). 1 In all 
these cases, Manva-brahman influence is very prominent. 
So also in the case of Saudasa-Aiksvaka 2 (54) , where, although 
only wife, Madayanti, is named in the story, it appears he 
had a harem which was in ' Vasistha's * charge during Ms 
exile (as with Trisanku) ; and this ' Vasistha ' had a hundred 
sons, and in addition was glad to obtain Saudasa-'s queen. 3 

The other cases of this interval are not definitely assign- 
able in time, and some of them might even belong to the 
next great period of polygamy (86-96). Thus Nlpa 4 (S. 
Pancala) is credited with a large number of sons, and was 
thus polygamous; but he may be placed anywhere between 
steps no. 65 and 80 ; but no cases of polygamy being definitely 
ascribable to the early Pancalas, while they are to the later 
Pancalas, the lower limit would be better. So also with 
Somaka (N. Pancala), 5 who had a large seraglio of about a 
hundred wives apparently, each of whom is said to have borne 
a son (100 altogether) after the sacrifice of the first child 
Jantu (who also is credited with 100 or 500 eons) ; Somaka 
and Jantu may be placed either immediately after Sudasa 
and Sahadeva, at step 70 (and 71), or immediately before 
Prsata, at step 90 (and 91), but rather at the lower limit 
for the same reason. It is to be noted that though the 
Pancalas (offshoots of the restored Pauravas) had intimate 

1 Vayu : 99, 16364; Matsya; 49, 45-46. 

2 Visnu: IV, 4, 19 38; Mbh. I, 17677 and 182; cf. also n. 3 


3 Saudasa's gr.-son Mulaka earned the appellation of * Narl-kavaca * 
., , (Vayu: 88, 17879; Brahmandla : III, 63, 179; Visnu: IV, 4, 

38); (though the story connecting him with Rama-Bhargava is 
anachronistic, the other statements about his living in the midst 
of many women may well have been authentic. Probably this is a 
very early instance of what was later on a common practice at 
courts, employing foreign female guards (e.g. with the Sindlm 
kings temp, the Epic, or with the Mauryas of the 4th cent. B.C.) ; 
possibly also, the so-called ' naked ' women surrounding Mulaka 
are really ' nagnas * or * maha-nagnis,' i.e., courtesans attendant 
on the king, which also is found in ancient Vedic (e.g. with the 
Vratya king) and subsequent court life ; in that case their import- 
ance would date from an earlier period than Lompada's, Similar 
* nari-kavaca * stories are also told of Haihaya princes who were 
beaten back; so that ( female-guards and courtesan-attendants * 
may go back to a still earlier period, being; apparently derived 
from the S. W. : a significant point. It is to be noted that 
female _ body-guards of ' apsaras ' like grace (who were also 
concubines) are well known to the brahman sacrificera in the 
Y. y. 

* Matsya: 49, 5253; Hariv. 20, 1060-62; etc.; it is however pos- 
sible that the ascription of a large number of sons is general 
here, and simply means that particulars about the family are 
not known for the next f-ew generations. 

5 Vayu : 99, 2035; Brahma: 13, 99101; Mbh. (Jantup. ) : III, 127 
128; Matsya; 49; Visnu: IV, 19, 18; H^riv. 32, 1793; all 
taken ^together, re Somaka and Jantu. (In Jantu' 3 case also 
there is the same possibility as in Nipa's; see above). 

( 209 X 

brahmanic relations from the beginning , yet it is only 
much later that they become ! priest-ridden,' ' as in the 
time of Somaka 1 or Brahmadatta 1 , while in the earlier genera- 
tions they themselves formed into fresh brahman groups from 
time to time and kept up an equality with brahmans proper 2 ; 
this apparently indicates the gradual and subsequent over- 
powering of t.e Aila element in the Pancala group by the 
brahman element ; and with this appears polygamy, whereas 
in the earlier Pancala period the only particulars available 
show a monogamy, that of the devoted and heroic Indrasena 
and Mudgala 3 (cir. 60) . The Paurava-Magadh'a cases of Vasn 
(who seems to have had 2 children by a different woman 4 
from his queen Gririka), and of his son Vrhadratha (who 
married the twin daughters of the Easi Mng) 5 , might be 
placed either .at steps 7879, or at 89 91 s but preferably 
at the latter period, with the Epic accounts 6 ; the noa-Aik 
and eastern connections of these princes are evident from 
every detail of traditions about them. 7 The ' Yrsni 9 who 
married two wives 8 , Gandhaii and Madri, may be put 
anywhere between steps 67 and 90, but better at the latter, 
for Vi'sni was quite a common name amongst the Yadavas, 
and Grandharis and Madris begin, to figure in the Maha- 
bharatan age specially. 9 But the ' Bhajamana ' (Sattvata- 
Yadava), who married th two daughters of Srnjaya 
(PancaTa) 10 , can be fixed in time 67. It is possible' that 
polygamy continued sporadically in the Tadava groups 
throughout the interval defined above; this would agree with 
the enormous expansion and multiplication of the Tadavas 
at the commencement of the nest period, as disclosed by 
detailed traditional accounts 11 ; the parallels between the 

t Clear from all the stories told about them. 

As in the time of the * Mudgala * groups of Ksatrlyan bramans. 

3 E.g. Ev. X, 102; Mbh. IV, 21, 651, (cf. Ill, 113, 10093); cl Hariv, 

32, 1781 (and other Pur. also) for Indrasena |Brahmi?tha mnst 
be itaken as an epithet of Mudgala). 

4 Called Adrika the 'nsh-apsaras ' in story ; i.e. a woman of the neigh- 

bouring Matsya clan. The two names however mean the same 
thing, and might refer to the same person. Cf. /.drika in Mbh. 
I, 63 ; and 123, 4817 ; and Girika in I. 63, 2367. 

5 Mbh. 275 (Kajasuyarambh. ) ; II, 17, 692 j 18, 726. 729, etc. 

6 As in Mbh. I, 63 and II, 1718, which do not show much contrac- 

tion of genealogy. 

7 Cf. the stories In I, 63 and II, 1ft 18 (Mbli.Jj Snttimati and 

Girivraja were in clearly non-Aila regions; thongli Yadava and 
Anava branches had settled there long ago. 

8 ' Krostu ' for ' Vrsni ' is an error ; the 2 wives are named ^ in all 

tfadava genealogies; Brahma: 14, 12; 16. 9 ff.; Hsriy. 35, 
19068; Matsya: 45, 1 2; Brahmanda : III, 71, 18 tt. 

9 E.g. in connection with the marriage-relations of tbe Kaaravas and 

Vrsnis just before the Bharata war. 

W In all" Yadava genealogies : Matsya : 44, 47-50 ; Vignu : IV, 13, 2: 
Brahman^: III, 71, 3-6; Hariv. 38, 1999-1003, Brahma; 15, 


H E e in Harivamsa, the wars, conquests, and expeditions; cf. Mbh. 
II, where the political situation clearly shows the central fact of 
Ya'dava expansion and rise to power. 

210 ). 

earlier Haifaaya-Yadava period and this Yrsrd-Yada/va one is 
significant. 1 The case of the Kaksivant, whom King Svanaya- 
Bhavayavya (and his queen Eomasa) gave a number of 
princesses in marriage; 2 is hardly assignable to any particular 
step; he may be t"he same as Dirghatamas' son (42), or a 
different person of the same line some 20 steps (65) or even 
50 steps lower (91) ; 3 ;what is definite here is that all the 
partis concerned in the case were Angirasas or allied to them, 
and the details have a distinct E. Anava character. 

At about 8 or 9 steps above the Bharata war, we come 
to a distinctly flourishing age of polygamy, illustrated in 
almost all the groups known to tradition; and during this 
century (which was also the close of the Bgvedic period), so 
full of personal details about great princes and connected 
brahmans, only two probable (and if definite, rather excep- 
tional) cases of monogamy 4 are . discoverable amidst the 
universal laxity of royal and priestly circles. 

Amonget the Aiksvakas of the time, the voluptuous Agni- 
varna (87) ; 5 fatally addicted to harem pleasures, and the weak 
prince Brhadvala (94) , whose fall at the Bharata battle was 
bewailed by his large number of wives, 6 are typical of the 
decline of Kosala. Amongst the East Anavas also, polygamy 
is now mentioned again: Vrhanmanas 7 (86) having two 
queens, from whose sons were derived two branch dynasties, 
and Kama 8 (94) marrying several wives ' according to the 
custom of the. family and the selections of his adoptive father 
Adhiratha.' Amongst the Pancalas (Pauravas) , the Southern 
section (of the above-mentioned Nipa's line) developed 
scandalous harems under priestly influence during Anuha 
and Brahmadatta '(86 87), as the Buddhist 9 as well as 

1 The working out of these parallels would lead to many interesting 

and suggestive results. 

2 Brhadd. : III, 141 fiO, etc, ; with Kv. I, 126. 

3 For the considerations that might help in locating him, vide Pargiter, 

AIHT., p. 223; the 3rd alternative is suggested by the occurrence 
of Sunaya .and Vitahavya as kings (90 ' and 91) in the Videha 
line, the mention of a Lomasa rsi in the same time in Mbh., 
and its mention of Cancla-Kausika of Gririvraja of the same 
period as a, son of Kaksivant. 

4 Balarama and Revati, and Drona and Krpi. 

5 Raghuvams'a : XIX, evidently based on common Puranic tradn. which 

it closely follows in many dynastic and personal details. 
G Mbh. (Stri-vil. ) : XI, 25, 715, 
7 Hariv. 31, 1705-6; Vayu : 99, 110 ff. ; Matsya ; 48, 105-8; (these 

were 2 sisters, dtrs. of a Vainateya, Saivya or Caidya king, prob, 

the last). 

Mbh. 569 (Bhagavadyana. ) : V, 143; cf. 'waves of Karna who 
would become widows * : Mbh. VIII, 87, (end). His son Vikarna 
also is stated to have had 100 sons : Hariv. 31, 1710. 

Gf . the several * harem '-scandals related of Ahese courts in the 
Jatakas, where the * Brahmadatta ' is evidently the Puranic one, 
the chief figures in the stories being the ministers Kan<Jan and 

( 211 ) 

some epic 1 traditions show, in spite of some other epic glosses 

which, connect these kings with the s Yoga ' cult.. 2 In the 
Northern section, Drupada, (93) followed up the polygamy of 
Somaka (probably to be placed at 90 within this period); 
one of his queens 3 (a sister-wife) is mentioned by name, but 
she was not the mother of Ersna and DhrstadyiiTrma, who may 
have been born of an abducted and widowed Eosala queen 
whom Drupada (re) married 4 ; Slkhandini too, was born of (or 
adopted by) ' ' the eldest queen ' ' of Drupada ; 5 and there 
were numerous other children by other wives", with whom 
Drupada was not satisfied; 6 he also gave away, along with 
Exsna, _ numbers of slave-girls to his son-in-law; 7 and in 
connection with the sex-fraud of Sikhandinl, who was 
married to the Dasama princess, a commission of courtesans 
from the Dasarna court was admitted into the Pancala harem 
to find out and report on the truth. 8 In the DvIMdlia 
section, Ugrayudha's (90) demanding the transference of the 
widowed Pauraya queen Satyavati to himself, 9 plainly 
indicates that with his conquests and sudden rise to power, 
he was making additions to the ' punarbhu * section of Ms 
seraglio, 10 and was simply following up his inevitable appro- 
priation of ifae North and South Pancala harems. u Amongst 
the Kurus (Pauravas) , the -cases of Pratipa (87, or 89?) and 
Santanu (90)' are negative and Inferable : the former could be 
polygamous but e restrained himself * (voluntarily or compul- 
sorily), 12 or his actual polygamy and transference of harem 
fco his successor Santanu may possibly have been amended in 
this way ; 13 the latter had two wives, but not at the same time 

1 Thus the episc statement that Brahmadatta's queen became a courtesan 
agrees fully with the Jataka statements regarding tlie amours 
of his mother and wives. 

This does nri mean anything ; * Yoga * cult was probably non- 
existent in 1100 B. 0. ; Besides e Yoga * is a cloak for many 
scandals; e.g. Karna was begotten by *Yoga'; etc.; vide n. 1. 

3 Prsati ; vide ante. She may have been the chief queen ; but vide n. 15, 

4 Aa the Jataka tradn. has it : cf. Jatakas (Oowell) : V, 225, etc. 

5 Mbh. 573 (Ambop. ) : V, 188; her " co-wives,'* for fear of whom 

she concealed gikhandinTs ses : ibid. 190-191. 

6 It is therefore that he wanted other offspring ; and as a result of Ms 

sacrifices and prayers, Draapad!, Bhrstadyumna and Sikhandin! 
were ' born/ 

7 Vide n. 5. p. 214. 

8 Mbh. V, 192. 

9 Hariv. 20, 10854112. . . 
10 Vide notice of such a section in connection with * widow-remarriage * 

(cf. his ancestor Ajamidha's 'punarbhu' wife). 

U Ugrayudha completely conquered N. and S. Paficalas, expelling and 
exterminating the princes. 

12 Details in Mbh. 162 (Sambhava. ) : I, 97; compulsion is more pro- 

bable, as his queen was a Saivya (cf. Jyamagiia's case). 

13 This would seem to be the real fact behind the story., sp. when 

compare* with the Jyamagha story, and taken with Satyavati s 
proposal of transferring the wives of Vicitravirya to Ins next 


( 212 ). 

apparently 1 ; yet Ms evident laxity 2 and the statement that his 
first wife had to employ all her arts of singing, dancing and 
coquetry to please him 3 >are indications of his harem-Jite. 
Pratrpa's other descendant, however, the Vahlika (Paurava) 
prince Bhurisravas, had a number of wives 4 ; and in the next 
generation, Vicitravlrya (91) 5 is a polygamist of the Agnivarna 
type (87) : his three wives are well-known to tradition, 6 but 
obviously he had many more in his short span of life ; and 
if Vidura's mother was really only an ' apsaropaxna dasi ^ and 
not a princess-wife of secondary rank, then it would indicate 
concubinage of slave-girls in his harem 8 ; but princess or slave- 
girl, she was a lawful ' ksetra ' of the king, 9 and no doubt 
others like her were. Of Dhrtarastra, only one wife anyone 
concubine are named, Gandhari and a Vaisya 10 maid-in-waiting 
(taken into favour 'during the former's pregnancy), 11 but 
clearly he ha'd many more to have 102 children ; in fact the 
alleged circumstances 12 of the birth of Yuyutsu and the 
other sons, the consistence and frequency with which 

1 Bhisma's mother was, however living all along, only she had left 

Santanu (or was left by Iiim) ; e.g. she educates Bhisrna for 
several years; takes part in the Amba episode, and is stated to 
have been living at the time of his death,- which is not very 

2 As shown in his adventures in the course of hunting expeditions and 

his treatment of Bhisma, 

3 Mbh. I, 98. 

4 Mbh. 619 (Stri-vil. ) : X, 24 (i) } 687; the chief wife seems to be 

mentioned in XV, 808. (PutradaTs'. ). 

5 Mbh. I, 102; cf. Vis. IV, 20, 10. 

6 He may hav-e had 4, if Amba had agreed to marriage with him. 

7 Mbh. I, 106, 4297. 

8 Slave-concubinage was not unknown to Kuru court circles before this; 

Kavasa-Ailusa (74) chaplain to the Kurus, was born of such a 
slave-jsirl, which was a ' subject of much comment * ; cf . the 
parallel case of Uddalaka son of the Pancala chaplain and a slave- 
girl, soon after this period (96/97). 

9 Mbh. I, I, 94; 63, 2441; 105, 4224; cf. Br. 13, 120-'l ; Hariv. 32. 

1825-'6 ; VS. 99, 235-'42 Mat. 50, 47. 

10 The notice of Vaisya concubines in this and some other cases, prob. 

points to- the existence of customs similar to those noted In 
Vats. Ka. Su't, fotr the post-Maurya period (i.e. presentation to 
the court of beautiful daughters and wives by the tenants). 

11 Mbh. I, 115. 

12 These are indicated in : Mbh. 180 (Sambhav. ) : I, 115 (within 

.a month the 100 sons and 1 dtp. of Dhrta. were born; during 
Gandh.'s pregnancy Dhrta. begot the Karana Yuyutsu on a 
Vaisya maid) ; cf . enumeration of Dhrta. 's sons in order of suc- 
cession of birth (so that the mothers were separate) : Mbh, I, 117. 
It is to be noted that at first only 11 names of these sons aro 
given (Mbh. I. 2446) ; then at the time of Draiu.'s svayamvara, 
23 are named : (Mbh. I, 6984 ff.) ; so that these sons were 
evidently not born at the same time and were born of diffeuawt 
mothers. Yuyutsu' s mother was something more than a con- 
cubine, almost a legal wife, as he is regarded as the continuer 
of DhTta.'s line and his ' pindadata ' : Mbte 577 (Bhisma- 
vadha. ) : VIS, 43, "*626; cf. other refs. to his 'birth from" the 
Vaisya wife in : Mbh. I, 63; II, 74, etc. 


Buryodhana is called ' Grandhari ' (or equivalents), 1 and 
JLXhrtarastra's statement that he was his son by his eldest 

queen, proves that there were other wedded wives and 
concubines too; Dhirtarastra is always iound attended by 
female slaves 2 who help him in his toilet and bath, and he 
consoles his son Duryodhana (envious of Pandava opulence) 
by pointing out what a choice lot of pleasnre-girls and wives 
he has placed at his disposal. 5 

Ehrtarastra's sons also were polygamous : besides the 
equipment already noted, Duryodhana had at least two wives, 
one distinguished as the mother of the heir-apparent 
Laksmana, 4 the other the Kalinga princess abducted from her 
* svayamvara ' 5 ; he had doubtless concubines amongst the 
female slaves of the palace, like his father, as is shown by. 
his indecent attitude towards Dranpadi in his * sabha/ whom 
he won into slavery and (then invited to be his concubine 6 ; 
several of Dnryodhana's brothers had separate palaces and 
establishments 7 (appropriated subsequently by the 4 younger 
Pandavas), which implies similar polygamy : thus, in the 
epic, Vidura is ascribed such a separate establishment and 
one wife (a Vrsni-Yadava princess) , 8 without any particulars, 
but the Jataka traditions supplements it by ascribing to 
him 9 palaces and numerous women, the full polygamous 
and luxurious royal style, equally wi|h the reigning Plndava 
sovereign. 9 Even Dhrtarastra's relatives by marriage, Subala 
of G-andhara (Druhyu _or Aiksvaka ?) and Jayadratha of 
Sindhu (Druhyu or W. Anava?), whose obscure families seem 
to have had no tradition of polygamy behind them, are found 
to be equally polygamous : thus Subala had a son who is 
distinguished from Sakuni, etc., by an added metronymic 1 ^; 
and Jayadratha, early in the life-time of Duhsala, apparently 
contracted another marriage, with a Salva princess, for he wae 
marching in wedding procession to her country, when lie 
met Draupadi 11 (a> sister-in-law) , and tried to abduc$ her to 

1 Called * Gandhan J 31 times In Mbh. ; cf . Sorensen : Index, p. 279. 

2 E.g. Mbh. XL, 12 IS; cf. sons of Sudra concubines compared with 

those of princesses like Gandhari : XI, 26 (beg). 

3 Mbh. H, 49; cf. young women of Dhrta.*s palace at the disposals of 

the Pandavas (guests) : II, 58; cf. also III, 259 ff. 

4 Mbh. XI : 17B, 511. 

5 Mbh. XII, 4. 

6 Mbh. II, 71. 

7 Mbh. XII, 44, 1517 fL . 

8 Mbh. I, 114, 4481-2; II, 78, 2568; 79; cf. I, 129 It is mute 

possible that Knnti lived with him as a wife, after her return 
from Satasrnea and during Pan^ava exile. 

9 Oowell : Jatakas, VI, 126 ff. ; 1,000 wives and 700 courtesans and 

slave-gi?ls in his palaces (p. 145). 
10 Mbh. VII, 49, 1933. 
SI Mbh. Ill, 264, 15576. 

( 214 ) 

make her his wife ;* his dead body, Is surrounded by his foreign 
slave-women, 2 who no doubt also kept him company in his 

Pandu. ateo, like his brothers, was polygamous. The 
similarity with Vicltrayirya's case> makes it probable that he 
Bad more than two wives (Kunti and Madri) ; the deer-story 
is clearly an invention to explain away the fact that Pandu 
was another victim to the prevalent debauchery of the courts 
of this period; Pandu 's own comparison of himself with 
Vicitravirya, and Kuntfs comparison with Vyusitasva, 3 both 
ehildtess, voluptuous and stricken with consumption in early 
youth, were thus quite appropriate. 

In spite of the polyandry, the Pandavas were as good 
potygamlsts, each one of them having other ' individual ' 
wives besides the common wife, who was the ' mahisL' 
Thus 4 Yudhisthira married a Saivya princess, Bhima the 
Kasi princess Balandhara (and the * Eaksasa ' princess 
Hidimba, in liis early youth), Arjuna the Eauravya-Naga 
princess TJIupi, the Manipura (Kalinga?) princess Citrangada, 
Ms cousin the Yrsni princess Subhadra (while he received 
offers of concubinage from one ' Urvasi,' and of marriage 
with Uttara-Vairati), Nakula the Cedi princess Karenumati, 
and Sahadeva the Madra princess Vijaya, the 3rd daughter 
of Jarasandha of Magadha, and the Yadava princess Bhanu- 
mati. Apart from these royal wives, the Pandavas -also 
appear to have had as ' full * harems as their predecessors 
and cousins at Hastinapura. They 'obtained slave-girls 5 
from Drupada along with their common wife ; at Indraprastha 
their palace was filled with large numbers of female slaves 
and attendants of all ranks and descriptions, amongst whom 
were bejewelled court dancing-girls and ' artistes, ' over whom 
Draupudf kept a vigilant eye, and every one of whom she 
knew by face and features, 6 but with whom nevertheless 
Draupadi and Subhadra, together with the princes, indulged 
in drunken, voluptuous merriment, in the royal summer 
resorts on the Yamuna. 7 Eunuchs also formed part of the 

1 Mbh. II, 267268. 

2 Mbh mS 22 v-, 62 Z\ cf ' tj ? e Sauvir ^ P rince Sa *Wa urged by his 

mother Vidula to enjoy sovereignty along with his Sauvira 
damsels and avoid the fate of being ruled by the Sindhu King 
and his Samdhava damsels : V, 134 

3 Mbh. I, 120-121 ; cf. Vis. IV, 20, 10. 

J 219-221; III, 22, 898; 45-46- 
70 A; IV 90, 3977-'83 ; SV, 1} 24; 25, 668/etc, etc.; 

; etc. 
' f " 196 (slave "S irls at P5fi cala court); vide n, 9 and 10, 

* Mbh/ m," 233-235. r 

1 Mbh. I, 

( 215 ) 

harem establishment (as they did at the Matsya court), 1 
for Draupad! herself had a favourite eunuch for her personal 
attendant 2 (which seems to be the original of the Jataka 
statements about her corruption with a deformed slave).. 3 
The ranks of slave concubines must have considerably swelled 
with the lavish nuptial presents from Subhadra's kinsmen , 
the "Vrsnis, of youthful South- Western slave beauties 4 (very 
probably imported in regular traffic by sea) ? 5 and with 
similar presents of Kairatiki (Tibeto-Burman ?) and llagadhi 
(Dravido-Eolarian?) slave-girls from subordinated or friendly 
chieftains. 6 Thus it was that the Pandavas could, it is said, 
provide 30 slave -girls for each of the 88,000 ' snataka ' 
brahmans resident in their capital', 7 and at the Kajasuya could 
give away as many to each of the assembled priesfe (with 
suitable temporary quarters, in the sacrificial area, for their 
new acquisitions), thus earning brahmanical gratitude and 
laudation. 8 This large and choice collection Yudhisthira lost 
to the Dhartarastras by gambling. 9 But even in the interval 
of exile, the Pandavas continued to be attended by the slave- 
girls obtained along with Draupadi, with whose restoration 
they too would be legally returned to them ; and they were 
recovered after their great victory, augmented with the 
only temporarily sent to the Pancala court for convenience. 10 
What the Pandavas lost to the Dhartarastras they must have 
appropriation of the several establishments of the defeated and 
killed princes; 11 and no doubt Bhima was then able to fulfil 
his promise on the battle-field to his charioteer Visoka or 
providing him with a hundred choice slave-wives and suitable 
establishment. 12 

1 Mbh. IV, 11. 

2 Mbh. W, 1. 

3 Cowell : Jat. V. 225 ff. 

4 Mbh. I, 223; Pan$avas received slave girls from Yadavas on the 

qccasion of Uttara's marriage, also : IV, 72. 

5 Thus the island possessions of the Yadavas in the Arabian Sea 

(Ratna-dvipa) "were noted for their -trade in gems ajad women : 
Hv. 95, 5233 ft. 

6 Mbh. II, 52, 1867; cf. XIV, 85; also VIII, 38, 1770. 

7 The figures are of course exaggerations; but that does not affect 

the point here. Mbh. IIS, 49; vide n. 5, p. 214. 

8 Mbh. II, 33; the brahmans who had come into close contact with 

the Paurava courts immediately before this, were also inclined 
towards polygamy or laxity amounting to it; thus the Vasithas 
Parasara (Sagara) and Krsna-Dvaipayana (90 and 92) had off- 
spring by more than one woman, some of them being Paurava 
princesses. (It is to be noted that the branmans connected 
with the Kuru, Pan^ava and Paficala courts in this period are 
mainly Vasisthas and Kasyapas, and partly Ingirasas, all of 
the inner Manva group by origin). 
9 Mbh. n, 60-61. 

10 Mbh. Ill, 23 (beg.). 

n Mbh. XM, 44, 1517 ff. 

S2 Mbh. VIII, 76. 


Amongst the families closely related to the Pandavas, 
polygamy is found with the Vasu-ite Pauravas, of G-edi, 
Magadha and Mafcsya, besides the Yadavas, who form a great 
polygainic group by themselves. The Caidya Sisupala (half- 
ifadava, half -Paurava) , in addition to his unspecified wife or 
wives, had .relations with Bh'adra-Vaisali 1 and Babhru- 
Yadava's wife 1 , and was betrothed to Bukmim whom he 
would have been glad to obtain even latter on\: his polygamy 
is thus inferable. The Magadha-Paurava cases of Yasu aad 
.Vrhadratha have already been noted as probably assignable 
to this period; no details are given about Jarasandha's own 
wife or wives, but two of his daughters were married to Kamsa- 
Yadava 2 ; and " magadhi " slave-girls having been at this 
time presented to the Pandava court, 3 they must also have 
been- part of (the equipment of the Magadhan court. About 
the Kasis of this time (intermarried with Kurus, Pandavas 
and Pancalas) no polygamies are mentioned, unless the 'twin 
wives of Vrhadratha and the cases of Ambika and Ambalika 
are taken as ^ evidence for the Kasi court itself; but apparently 
the Kasi prince who was a contemporary of Dhrstadyumna 
in iiis youth, was polygamous. 4 Particulars about tlie Matsyas 
are much fuller : Virata had at least two chief wives, Sudesna 
of Kekaya, and the sister of ' Klcaka,' 5 and he had oFviou's'ly 
a big ' harem ' ; his young son Uttara is already a gay reveller 
in .the company of numerous women. 6 The court of Virata 
is^of the same type as that of the Pandavas, whom indeed he 
tries to. follow and emulate in many respects : there is the 
same fondness for gambling, 7 and employment of eunuchs in 
the ^ personal service of the princesses, 5 the same normal con- 
cubinage of the female attendants of the palace, 9 and volup- 
tuous ha-rem-life of the princes amongst multitudes of women 
with song, and dance, and wine 10 ; all illustrated in the Indra- 
prastha and Hastinapura courts ; in one respect apparently 
the Matey a court made an improvement, in special arrange- 
ments for dances as a court pastime ; but probably this wa>s 

1 Mbh. II, 45; Hv. 117. 

2 Mbh. II 3 14, 594; Hariv. 91, 4955-'61. 

3 Mbh. II, Rajasu. & Dyiita. 

4 Mbh. VJCI, 10 3 364. 

5 Mbh IV, 249. 432. 562. etc,; 18, 529 (Kicaka a ^yala; hence, 

6 Mbh iY a ^ a< ii queen ' a Kicald ' for Sude?1 ^ was a Kaike yi). 

7 Of Virata, who was glad to employ the supposed expert games- 

master of Yudhithira. 
Of V?hannala-Arjuna as dancing-master and personal companion of 

Uttara, who almost fell in love with him. 
Of. the cool presumption of the king's brother-in-law, as well as 

of the queen, with regard to Draupadi the Sairandhri. 
Hi. g , the life led by the effeminate Uttara even in times of great 
tm, dai ? er > and b y th martial Kicaka addicted to wine -and women. 
11 Inus there was a special dancing-hall, apart from the usual 
sabna , where oowt ladies had free access. 



( 217 ) 

imitation of the adjacent Yadava courts of the South- West, 
where, from much earlier ages, whole royal families trained 
themselves 1 as expert singers and dancers, with occasional 
resultant scandals. 

Amongst the Yadavas of the same period, full-fledged 
polygamy is almost the general rule with the confederate clans 
of Dvaravati (at first of Mathura) headed by the Vrsnis, 
and they were the representative Yadavas of this age, just 
as the Caitrarathas and Haihayas were of an earlier one ; 
instances of polygamy sometimes occur also amongst other 
contemporary branches of the Yadavas. 

Taking the Vrsnis, amongst whom Krsna was born, we 
find instances of polygamy in all its four main branches. The 
group indeed began with the polygamy of ' Vrsni/ who, as 
already noted, may be placed within this period as well (at 
88/89). In the first branch, Satrajit '(92/93) married 10 
sisters, 2 and of his daughters three were given in marriage to 
their cousin Ersna. 3 In the second^ Atriira had at least 
three wives named in the Mists, 4 besides others who took part 
in Eaivataka and other festivities, 5 and who all retired to a 
Himalayan hermitage after his death. 5 In the fourth, no 
definite particulars are available, but both Sini and Satyaki- 
Yuyudhana were * bride-abductors ' of the Krsna type, 7 and 
their polygamy is quite likely ; the Vedie story about Asanga- 
Yadava, 8 which has a marked ' h'arem * character, might very 
well refer to Satrajit 's son or grandson Asanga. 9 ; and another 
Asanga, a son or grandson of SatyaM, was apparently settled 
by Arjuna on the Sarasvati, with a part of the remnant of the 
Yadava harems. 10 Eor the third branch! more details are 
forthcoming, being Krsna's own family. Devamidhusa 

1 E.g. in the families of Durjaya, Tittiri, Revafca, and amongst the 

Vrsms of Krsna's time. IJttara's training in dancing was not 
exceptional amongst the Pauravas; Santanu's first qneen ^was 
an expert dancer and singer; and the etmuch whom Arjiina 
personated, must have done similar service to Draupadi. 

2 Va. 96, 53; Br. 16, 45 j Hv. 39, 2076 j cf. Mat. 45, 119. 

3 Br. 16, 478; Hv. 39, 20789; gr. dtrs. : Mat. 45 ? 1 21 : cf. VS. 

4 Sutanu Augrasem : Br. 16, 55 ; 14, 8 ; 11 ; Bd. Ill, 71, 113 ; Hv. 

35, 1919; 39, 2086; Batna gaibya : Mat, 45, 2732; AsVini : 


5 Mbh I," 219, and Pur. ace. of Eaivataka and Frabhasa festivities; 

cf land and sea sports at Dvaraka where Bhanunati is abducted . 
Hv. 147149. 

5 3U- Mbh. VII, 144, 6032-43; SatyaH : VII, 10 p; | 
tions by Krna, cf. Mbh. II, 45, 1574-7; III, 12, 575; 
158, 535 ft. ] 5364; 48, 1881 fE; VIE, 11, 391 fLj .etc 

8 Ev. VHI, 1; cf. Mait. Sam. Ill, 1, 9; Brhadd. H, 8, 3; VI, 41; 

9 Padma.' V, 13, 94-6; cf. Mat, 45, 
10 Mbh. XVI, 7, 245 S. 

( 218 ) 

(89/90) probably had two wives, one an Asmaki, 1 and the 
other a daughter of Aryaka, the Naga chieftain (on 
the Upper Ganges) 2 ; his son Sura (91/92) also, is ascribed 
two queens. 3 In the next generation, Vasudeva (92/93) is 
a prolific polygainist : amongst his 20 wives who bore him 
children, were his 7 cousins, daughters of Devaka, 7 Naga 
princesses, and 2 maids-in-waiting, besides Eohini-Pauravi 
(sister of Vahlika), Bhadra-Vaisali, and Madira (which three, 
with DevakI of the first group 5 ascended his funeral pyre). 4 
His eldest son Rama is ascribed only one wife, Eevati, 5 but 
his monogamy was apparently not puritanic, as he took a 
prominent part in the Yadava drunken orgies, 5 and is said to 
Have indulged in wine and women along with Krsna, while 
on a military expedition to the G-omanta hills and Earavirapura 
(in lower Deccan). 5 But Krsna followed his father and 
went far beyond him, being in fact the greatest polygamist 
of his age. Much 'detail is available 6 about the wives of Krsna, 
even a summary of which would be lengthy; but they "may 
be classified as (i) cousins or near relatives given in mar- 
riage: e.g., Satya-Satrajiti and her sisters, Akrura's sister, 
etc., (11) external princesses (some of whom were related), 

1 Hariv. 35, 1922 ff. ; Bd. HE, 71, 1456. 

2 Mbh. I, 128, 5026; cf. V. 103, 3635; etc. 

3 Va. 96, 14344 (prob. 3 wives); etc.; cf. Hv. 95 : 525152, where 

Vasudeva's f. is called Vasu (for Sura) and is ascribed 3 wives. 

4 Va. 96, 12931; 149166; Bel. Ill, 71, 145163; Mat, 44, 723; 

46, 1121 ; etc. ; cf . Mbh. II, 1570 ; VII, 144, 6032 ft, ; XVI. 7, 
224253. ' 

* (a) Va. 86, 26-29; 88, 14; Bd. Ill, 61, 1924; 63, 13; Br. 
7. 30-41; Hv. 10, 64411, 657; Vis. IV, 1, 2037; 2, 12: 
cf. Mbh. I, 219, 7912. 

(5) e.g., in the Raivataka and Pirabhasa festivities. 
(c) Hv. 98, 5405-'39. 

6 Jambavati, Satyabhama & Akrur-a's sister : Va. 96, 20 98, Br, 16. 
1245; 17, 140. Knkmim : Hariv. 161, 91346; 117118. 
Gandhari: Hv. 161, 9147^-8; 174, 9797. Kalindi-Mitravinda 
Satya-NagnajitT, Eohinl-Jambavati, Suslla-Madrr, Satya-SatrajHi, 
Laksmina-Jalahasini, Tanvi-gaibya : Hv. 118, 67006706. 15 
wives detailed : Mat. 47, 1323. Satrajit's 3 dtrs., Satya, Vratini, 
Tapasvini (or var. lee.) : Br. 16, 478; Hv. 39, 20789; Mat. 
45, 1921; Bd. Ill, 71, 242264. For Mbh. refs. to brides 
abducted by Krsna, vide n. 7, p. 217. 16,000 wives and prin- 
cipal wives (widows, self-immolating, retiring or captured) : 
Mbh. XVI, 5; 7; for 16,000 (captured from Naraka), cf. Mbh. 
I, 2789; V, 130; 158, 5353 ff. ; XIII, 15 (end); 160, 7422 f . ; 
XVIIU, 5, 1713; cf. Hv. 122124. For Mbh. refs. re 
Satyabhama, vide III, 183, 12567; 233235; IV, 9, 262; 
re Rukmini: II, 2; Xllli, 617; 621 fE.; 139 (beg); 149 
(beg.) ; 160, 7416 fi. ; cf . Hv. 14406 if. 5 re Jambavati : 
m, 670; 10271; XIII, 1415. Mbh. XVI supplies Haimavati 
as the name of another chief wife. Cf. generally, Br. 199, 201, 
202, 205, 212; Hv, 37, 194768; 124, 69626999; 162; Padma. 
V, 123; 170. For shepherdesses as wives or concubines, cf. Br ( 
184: 189; 193; Hv. 7& 79; cf. Mbh. II, ^91, r 

( 219 ) 

either ordinarily married, or abducted b~y force (in ' svayam- 
varas ' or other circumstances involving fighting) : e.g. , 
(a) Sufila-Madri, etc., (b) Eukmini-VaidarbhT, Satya- 
Nagnajiti, Jambavati, etc., the first three being related, 
(iii) sundry others, numbering altogether several thousands, 
1 married ' or in concubinage, consisting of other harems 
transferred by conquest, of ' artistes/ slaves, and probably 
even shepherdesses in state employ. 

Thus the entire harem of Naraka of Pragjyotisa was 
transferred to Dvaravati ; l its numerical strength is usually put 
at 16,000 and only once at 1,000. 2 Elsewhere the grand 
total of Krsna's 'wives' is put at 60,000 ; 3 probably this is 
intended to include the above 16,000, as well as the 16,000 
shepherdesses he dallied with-, 4 the remaining number being 
slaves ; but possibly this detail of 16,000 gopinis * may have 
arisen out of the more authentic one about Naraka's transferred 
harem. Or this also may have been traeT: for one thing, the 
theory that Ersna's ' gopa-lila ' is an Abhira accretion of 
later times, is untenable, as the Abhiras are traditionally 
connected with the Tadavas and the whole of the South- 
West of India/, from the time of Haryasva and Madhu (i.e., 
bet. 27 and 63) ? a_nd formed the subject population under the 
Yadavas; 5 other Abhiras again are mentioned as attacking 
the fallen Yadavas ; 6 they may have been rebel tribesmen 
having their day after all the injury they had suffered from 
Krsna, probably even, thus reclaiming their women, abducted 
by the rulers from time to time; for another, concubines 
of the Vaisya class were customarily taken into the Paurava 
and Yadava harems of this time (e.g., with Dhrtarastra and 
yasudeva) , 7 as also much 1 earlier, in iiksvaka and Vaisaleya 
ones probably (e.g., Dasaratha's queen Sumitra was a 

1 Vide ante, sec. re widow-remarriage; Hv. 123, 6934- f 61; Br. 202; 

etp. ; cf . n. 2 below. 

2 16000 : Mbh. V, 158, 5353 If. ; 1000 : Mbh. V, 130. 

3 Mbh. XIII, 160 (end). Here again all the figures are obvious 

exaggerations standing for ' a large number/ 

4 Vide n. 6, p. 218, re gopinis. 

5 E.g., the Abhira settlements (' vraja ') at G-oknia, VrndaYana, etc., 

under pastoral chiefs like Nanda,, ?nd owing allegiance to Yadava 
princes; vide n. 6, p. 218, re gopinis. Abhiras under Haryasva 
and Madhu : Hv. 51425167. Vide Sor. Index, s. v. ; bands of 
Abhiras of the S. W. were also snbj. to the Pan$avas and 

6 Mbh. XVS, 7 5 Br. 210212; Vis. V, 3738; Pad. VI, 279, 56 f.; 

Ag. 15 ; etc. 

7 Dhrta.^Mbh. I, 63; 115, 4522, etc.; II, 74; etc. Vasndeva : Mai 

* 46, lK-21 ; cf . B^. IEI, 71, 163. With Vaisya concubines may 
be compared Siidra concubines at the same two courts, e.g., 
with Devaka & Vicitewlrya : Mbh. I, 114, 4480 ; 109, 4361 j etc. 

( 220 ), 

daughter of such a Vaisya, and Nabhaga abducted a Vaisya 
tenant's daughter) 1 ; the traditional license of princes and 
high officials with regard to state-shepherdesses and other 
work-women is noted by Vatsyayana 2 ; and in Krsna' s time 
the Yadava clans rose to opulence largely by such state- 
pastures (cf. the ' vraja ' settlement under Nanda, and the 
details of Kalayavana' s birth , both showing connection of 
princes with the c gopims '). 3 A ' thousand ' of Naraka's 
women are stated to have been e ' married * ' to Krsna 4 ; thus 
the major portion of that lot of 16,000 remained only con- 
cubines ; to one of these sections, probably to the former and 
more favoured one, must have belonged those select ' gan- 
dharva * maidens (i.e., dancers and singers) whom Naraka 
had collected in his specially built hill-station of Maniparvata. 5 
Krsna' s harem evidently included slave-beauties also, 6 num- 
bers of whom he sent to the Pandava court, and assigned to 
the rsi Durvasas 7 when he lived in his p'alace as a guest. It is 
noteworthy in this connection, that a kindred Yadava clan 
(under Harita) is stated to have colonized some generations 
back, an island Eatna-dvipa in the South- Western Sea, which 
was specially noted for its pearls and beautiful women, and 
peopled by ' nisada ' sea-faring merchants and * madgura ' 
pearl fishers 8 ; the reference is obviously to the pearl and slave 
trade of the Arabian Sea, a. very ancient one indeed, going 
back to the second millenium B.C. and the pre-Aryan civili- 
zations of South- Western India.. It is rather striking that 
the number 16,000 is attributed to the harems of other 
princes of the Epic age also, by the Jataka ttraditiotf; and 

1 Das'. : ' Snmitra Vamadevasya Vabhuva Karani-Suta * Nabh. : 

Mark. 113 ff. (where the rule is mentioned that the 1st wife' of 
a prince must be Ksatriyan, and other wives may belong to other 

2 Vat. Ka Sut.^V, 5. So the number 16000 is not very much of an 

exaggeration; it would simply mean that all such women were 
exposed to the license of it-he Yadava rulers 


4 Mbh. V, 130. But (perhaps in a general way) 16,000 ' widows ' are 

spoken of In Mbh. XVI, 5, 144, and 16,000 " parigrahah ' in I, 
2789 1 XVI, 138 5 XVIII, 171-73; cf. XIII, 160, 7422 ff." 

5 Hv. 121 ,- etc. ; vide n. 1, p. 219. 

6 Vasndeva had at least 2 such ' wives ' (vide ante) ; cf . Mbh. IV, 72. 

7 Not to be confused with the earlier Durvasas; Mbh. XIII, 160, 

8 7416 f . This as practically the only instance of brahmanical connec- 

tion with Yadava polygamy, which for this period, seems to 
have been rather a result of military power and opulence by 
trade, etc., as also with the Haihayas mainly. The Yadavas of 
this age are in fact the least brahmanical of the ruling families. 

9 . Hy. 95, 5233 ft JThe recent Punjab and Sindh excavations prove 

that the Arabian Sea trade was flourishing in S. W. Indian ports 
in the 3rd mill-enium B. 0., and was pre- Aryan. 

. _ 

Cowell: Jat. I 231 (84,000 w., of Sudarsana of EtfSvati or Kusi- 
-S ag n ara ) V 264 ( 16 > w -> f Brahmadatta of Varanasi); 289 (of 
?-?4T V5ir " B y un g eslfc of 1( 50 sons, chosen King of Gandhara). 
JO, 222 (of a Vddeha King); 246 (of Br. of Var,), IV, 78 

( 221 ) 

ancient Persian harems were equally large; 1 after all it may; 
not have been altogether an exaggeration, and is intelligible 
when some of the astounding traditional royal customs and 
privileges noticed by Vatsyayana 2 for a subsequent period 
are taken into consideration. 

Of Krsna's descendants, his sons Pradyumna* and Samba 
had several wives ; those of the former were 3 : Subhangi 
Vaidarbhi, his maternal uncle's daughter, Prabhavati 
daughter of Vajranabha, and Mayavati, widow of Sambara 
(the last two being obtained by romantic adventures and 
force) ; those of the latter 4 were: Candra-vati, niece of the 
same Vajranabha , and Duryodhana's daughter (both obtained 
by adventure and fighting); besides he had intrigues in his 
father's unwieldy harem 5 (for which he was cursed along 
with the guilty wives) ; and Pradyumna .also must have had 
a richer harem-life than that indicated by his 3 wives, to earn 
his later estimation as an incarnation of Kama. 6 EJrsna's grand- 
son Aniruddha 7 also -had two wives, Bukmavati of Vidarbha, 
also his maternal uncle's daughter, and tTsa, "daughter of 
Bana of Sonitapura (the latter obtained b'y adventure and fight- 
ing, as in the cases of Pradyumna and Samba) ; and Knirad- 
dha's son Vajra inherited a considerable harem from his 

(of Dasaratha of Kosala); 200 (of Snrnci or Ruci of Mithilaj 
ace. to custom of having 4,000 w. from 4 diff. sections of sub- 
jects) ; 285, 288 (of Naga King Gampeya of Campa). V, 2 (of a 
chaplain of the Var. King); 97-'8 (of SUtasoma of Var.; 16,000 
-1-700 principal wives : cf . VI, 20-31, of Maha-Janaka, VI, 75 , 
of Canda of Var. 5 --of Vidura of Indraprasths, etc.) ; 133 (16,000 
w. passed, each one as wife, from the father Arindama of Magadha 
to his son and succ. Dirghavahu) ; 141 (of Okkaka *. e. 'Aiksvaka' 
of Kusavati or Kus'magara in Malla Kingdom; sent oat in 
batches for open license -in the streets, along with the chief queen, 
so that some one of them might conceive and bring an heir to 
the king who looks on). VI, 1 (of a Kasiraja. the chief queen 
being a Madri, Canda); 115 (of Angati of MitMla); 128 (of 
Dhananjaya Kanravya of Indraprastha) ; 249 (of Sanjaya of Sivi, 
Madri Phusati being chief queen); 252 (of his son Vessantam, 
obtained along with his chief queen, another Madri). 

1 Sykes : Hist. Pers. I, 507 (12,000 w. of Khusru Parviz, Shurin being 

chief queen). Gf. hundreds of wives of AchameBlan Kings, e.g. 
Artaxerxes Mnemon } Sykes : Hist. Pers. I, 186-*7 ; 247 ; cf . also 
3,000 desc. of Fath AH Shah at his d. : I, 183. 

2 Vats. Ka. Sut. V, 5. , 

3 gubhangi Vaidarbhi : Hv. 119, 67Q7-'26j Prabhavatl, dtr. of Vajra- 

nabha : Hv. 149 3 8474 ; 150456 ; Mayavati, w. of gambara : Hv. 
163-167; BP. 200-201. _ . 

4 Duryo.'s dtr. : Hv. 120, 6765-86; Br. 208; Gandiavati, niece of 

Vajranabha : Hv. 150-156. 

I Ma? 4" 1-21 \ cf C *Pradynmna and the dcrs. of Brahmadatta a YSdava 

prieSk Hv. 142, 7993-8053. __ 

7 Rukmavati : Hv. 119, 67Q7-'26 ; Br. 201 ; Usa, dtr. of Bana : Hv. 
- 175-190; Br. 205-207. 


predecessors, with which he was removed to Indraprastha 1 
(with its Pandava harem traditions). 

For the other groups of the confederacy^ the details are 
much fewer; but the polygamy of their chief members is 
often alluded to in general terms ; thus the ' Andhaka Maha- 
bhoja/ TJgrasena and Ahuka( the Kukura leaders), had all 
their hundreds of wives, accompanying them in the Eaivataka 
and Prabhasa revelries. 2 Ugrasena's Vaidarbhi wife Padma- 
vati's 3 illegitimate son Kamsa had two wives (daughters of 
Jarasandha), as already noted. Devaka had, besides his 
chief wife or wives, at least one Vaisya or Sudra concu- 
bine, 4 whose daughter was married to Vidura-Kaurava. 
EJrtavarman's section was also apparently polygamous, for his 
descendant and heir inherited a harem with which he was 

settled at Marttikavata. 5 

Amongst the other Yadava branches, no polygamies are 
ascribed to the Vidarbhas, consistently with the Vidarbhan 
tendency towards monogamy. But the Salva (Marttikavata- 
Bhoja) prince, Mitrasaha, a contemporary of, Vasudeva 
(92/93) , had two wives (whose sons were the famous Hamsa 
and Dimbhaka or Nimi) ; 6 and probably the Euntis were poly- 
gamous, if the l TJrvasi s episode of epic tradition may be 
referred to the court of Purujit-Kuntibhoja. 7 

After the polygamies of the century and a quarter 
described above , there was apparently a natural temporary 
reaction in the Paurava court, 8 no doubt aided by times of 
great ^stress and reverses, 9 clearly indicated even in the meagre 
tradition of the period. Thus only one wife (the princess 
of Madra) 1 * is ascribed to Pariksit II (96) , u though he may; 
have inherited the Pandava harem in the same way as his 
contemporary and related Yadava princes. 12 His son Jana- 

a Mbh. XVI, 7; vide n. 6, p. 219. 

2 In all epic and Bur. accounts of Raivataka and Prabhasa or Dvaraka 


3 Padma. II, 48-51. The statement that she mistook another person 

Gobhila for her husband Ugrasena, shows that she was only one 
of a large group of wives; cf. the case of Bhadra-Vaisall, who 
being one of many wives could be similarly beguiled (by 

4 Vide n. 7, p. 219. 

5 Mbh. XVI, 7, 245. 

6 Hv. 295, 15387- '405. 

* Vide ante, pp. 139141. 

8 The Yadava (Indraprastha) and Pancala courts very soon disappeared ; 
and no personal details are available for Kasi, Kosala, etc., for 
this period. Vide Pargiter AIHT. pp. 284-'5. 

s Involving the retreat and amalgamation of the Kuru-Pancalas, dis- 
appearance of Yadava and W. Anava kingdoms, and considerable 
icontraction of the Epic horizon. Vide Pargiter AIHT. pp. 284-*5. 

10 Madravati" : in all dynastic accounts. 

11 Abhimanyu had also only on-e wife, but he was killejftt 16. 
~& Now on his Western frontier. 

( 223 

mejaya HI and his wife Vapustama, the Easi princess, are 
stated to have been particularly devoted to each other and 
to have led a monogamistic life.; 1 this agrees with what is 
said about Janamejaya's indignation at Ms queen's defilement 
through Yajurvedic ritual, and his reacceptanee of her ; 2 it 
is probable that Janamejaya led a reaction against the preva- 
lent corrupt practices of the priesthood 3 and the court, witE 
some amount of success, which however was short-lived 
owing to brahman hostility. 4 His son, Satanika (98) also is 
ascribed one wife, a Vaddehi ; 5 but Svetakarna 6 (100 ?) prob- 
ably had an usual harem, as he is said to have left for the 
woods owing to childlessness, before he met en route 1 the 
Yadava princess Malini (of Krtavarman's family) who bore 
him a son. 8 

The Videha dynasty, remarkably free from indications of 
corruptions, shows the first clear instance of it in the time 
of Janaka-TIgrasena-Puskaramalin (98) or Ms successor, at 
whose court Astavakra was entertained and tempted by courte- 
sans or slave-girls. 9 But the notices of the subsequent 
Janakas like Dharmadhvaja (with his spiritual consort 
Sulabha, 10 the bhiksuni ' daughter of a 'rajarsi') or Jana- 
deva (connected with P-anca&kha) ? 10 show _that tjns fall in 
moral standard was temporary or intermittent. 

Apart from the already noted harem-inEeritance of the 

three surviving Yadava princes, tha only other pertinent 

I In all dynastic accounts. 

3 Hv. 195, 11232-11278; cf. J.'s discussion with Vyasa re evils of 

horse-sacrifice : Hv. 192, 11092. fi. j cf. also Mat. 50 ; 57-65 and VS. 

99, 250-6. 

3 For which cf. also the traditions connecting then with laxity and 

prostitution; vide infra.; cf. also laxity of Buci, w. of Deva- 
s'arman, a contemp. of Janamejaya : Mob. XIII, 40 ff. 

4 The general tendency -of the priesthood and the Courts to lasity and 

polygamy showed no signs of abating in this post-Bharata period, 
and the two groups went oa corrupting one another till the rise 
of Buddhism, which for some time kept thesa in check. 

5 E.o-. Mbh. I a 95, 3838. But another son Candrapi^a (unless lie is 

& the same as gatanika) had '100 eons 5 (called Janamejaya Ksatras) : 
Hv. 191, 11065-'7. 

6 Hv. 191, 11068-11072; Br. 13, 124-132. 

f This probably shows abandonment of Marttakavata also (like iadra- 
prastha and Hastinapura) and pushing back of ifae Bt ^J as on tbe 
feaus'amM side. , . . 

8 About 3 centuries later the famous Udayana of this line nad a inu- 

6 Mbfa fle inf 132^0599-134, 10690; XIII, 19, 1390 f.; cf Earn. VI, 
121 16 Br 212, 72 f . ; Suka, son of Vyasa, was similarly tempted 
either at this or at another somewhat earlier Janaka's court : 
Mbh. XII, 326; cf. another earlier Janaka, temp. Pratardana, 
cheering his troops with prospect, of numerous Gandharva girls 


TOB ^ jha the' same as the'Sulabha Maitreyi' of the 

was honoured as a ?i ? Vide n. 9, p. 34. 

( 224 ) 

Tadava detail for this period, is the fate of Krsna's harem: 
Some of Ms chief wives resorted to * sahamarana and some 
others retired into forest life ; but the great bulk of his famous 
' 1C 000 ' were carried off by the Abhlras, 1 and are stated 
to have been subsequently reduced to prostitution, in which 
profession, they were confirmed and instructed by ^ l)albHya- 
Caikitaneya in the same manner, it is said, as the Danaya 
women of yore were reduced to and instructed in that protes- 
sion by ' Indra ' after the ' Danava ' defeats ; and several 
Puranas profess to give the substance of that instruction^ ; 
they give two explanations of this fall, Krsna's curse on his 
wives for their infidelity, and resultant ravishment by the 
Sbhiras, and Astavakra/s cursing 4 a host of ' apsarases 
( = Bjsna's wives) to become courtesans. Several important 
probabilities are suggested by these statements: firstly, that 
6 organized' prostitution (under royal patronage) may have 
arisen as early as the time of the first Aila- Mngs, 5 under 
conditions associated with conquests and subjection; secondly, 
that the destructive Yadava wars 6 of the Epic Age produced 
rejtetition of similar conditions and results, on a large scale; 
thirdly, that between the harems of the time and courtesans 
the connection was very intimate, the former leading to the 
latter by degeneration, the latter to the former by sublima- 
tion/ so that royal polygamy often implied patronage of 

1 Br. 212 S cf. Mat. 70; Mbh. XVI, 5, 135, 144; 7, 223270. 

2 Mat. 70 j etc. 

s Even the " courtesan's art. " had its ' rsi ' .and * sutrakara * teachers : 
e.g. Svetaketu, Pancala-Babhravya, Dattaka and Vatsyayana; 
apparently Dalbhya was Svetaketu's immediate predecessor in 
this respect (within 3 gens, of each other) ; some steps before 
him, another Paficala-Babhravya of Brahmadatta's court, has a 
similar reputation in tradition. 

4 An anachronism by three generations. 

5 Some of the early ~Aila kings were actually ' Indras/ ace. to tradition, 

and many others ,took part in the ' Devasura J wars, as compared 
with only one early Manva king, Kakutstha, who helped an 
c Indra,' prob. his contemp. Aila, Nahusa, whose son married his 
daughter. Nahusa is stated to have taken a licentious turn after 
his victories, and Yayati's unwieldy harem of Bhargava-Diinava 
slave-girls, and his temptings by an ' Indra/ are well-known. 
The alleged develo ament of prostitution by ' Indra's ' victories 
may thus refer to t'ais period, if there is any traditional basifl for 
the statement ; cf. wars temp. Baji and Yayati : Mat. 24, 37 f . j 
25, 8 f. j also 12 other wars : 47, 41-241. 

6 A comparative study of the whole body of Epic-Puranic tradition will 

give a clear impression that the ' Epic ' age was a * Yadava ' 
age, >and the Bharata battle was only one incident in a series of 
destructive Yadava wars, singled out for epic treatment. The 
Puranas would seem to be quite right in their estimation of 
Krsna as the central figure of the age; it is remarkable that one 
Plir. refers to an earlier rather different ace. of Krsna's exploits 
' that reads like a smasana ' : it explains the Puranic conception 
that he was t born to lighten the burdens of the ^jlftrth. 1 
Z These processes are illustrated in the courts of Brahmidatta of iPafl- 
cala (87) and his father (86), in Krsna's Harem (94), and in that 
of Arjuna's 'real* father (93) (whoever he may have been). 

I 225 ) 

prostitution; 1 fourthly, that the brahmans, were chiefly 
instrumental 2 in fostering and sanctioning the profession^ as 
is shown by the prominent part played in these developments 
by Dalbhya-Calkitaneya and Astavakra (nephew of the 
notorious Svetafeetu, connected with brahmanical laxity, and 
a teacher of Erotics inclusive of Prostitution) , all of whom s 
remarkably enough, were Kuru-Pancala braHmans, together 
with the two Pan^ala-Babhravyas 3 similarly associated in 

1 Such patronage is indeed illustrated in very much earlier conrts 3 as 

already noted. 

2 So also in earlier periods^ the connection of Sukra-BAaryata with 

Yayati's slave-girls, o DM-Atreya with his host of pleasure- 
girls (and with his contemporary Kausika's addiction to courte- 
sans), of Bsyasriiga-^Qsya^a withAnga courtesans, etc., are more 
than accidental ; cf . Maitreya living with a courtesan at V&ranasi. 
temp, Vyasa who stays with them as a gratified guest : Mbh. XII. 
120-123: cf also Narada and the -prostitute Pancacuda (XITI, 38 ? 
2203 ff.), or the rsi Bodhya and ]?ingala (XIII, 178 ; XH, 174). 

3 They are most probaoly not the same . 








Page 1, 

lino 2 

o text, /or 

Bgveda *e^ 





JJ ^ 11 33 





j* ? 

)i J) 33 !) 





,3 15 

>, tCXfc ,, 


c gotha* 



?1 33 JJ 

33 33 





^ n. 1 

music- and >, 

muaic- and 


9 9 *^ 

3J >3 5 j, 

Uyhad, Upan. ,, 

Brhad. Upan., 



3, a 

3? 33 V 33 


V. L, 





13 33 " 33 

Upon. ? , 




33 1 

33 J3 ^ 35 

Sankh./ 3 , 




? 3 2 

37 33 I" J? 

Javadratha ,> 




33 1 

3) ?3 * 33 





33 2 

33 33 3J 33 

3J }f 




,3 2 


33 33 33 

T. 22 

T. 22 




33 6 ,, 

Sam. j? 


? 3 


33 7 

33 33 8 ,, 

Ajanta , f 




33 2 

33 33 * 33 

'itawa' tt 



3, 15 

n )'''}) 33 

c istaka' 




,3 20 

)> JJ 33 J3 

*it((J)a* ?J 




33 21 

Ji 33 3' 

'ita-ppu' ;i 




33 22 

3 J 33 1) 33 






3 ' >? 33 33 





33 1 

33 3? *^ 3> 

Sam. ; 


; ' 



33 J' 33 33 

73 )J 




33 17 

)> 3,1 33 )> 





3, 6 

>3 53 5 }) 

'rupam sukrtam' 

'rupam sukrtam* 

3 J 


,3 1 

33 3, 10 

sthuna 3 : 




57 text ., 

f methi'sl J3 




! 3 

33 33 3? 





33 6 

3J 33 33 



J J 


33 1 

,, n. o ,j 

Sam. j? 




33 3 

3? 33 ' 33 




33 12 

33 33 33 33 

Orissa) ,, 





text ,, 

flames/ ?j 


5 t 

5 3 

v 1 

?) n. i 53 

'sagad' ,, 

OQ fQ fs 

jjjpj V* jyj \jftp 

f * 

if if 

3 } 

3, 3 

33 33 3J 33 

cadinti J; 




3, 4 

33 S3 33 33 

e ^agad' 

's r aga<f 

1 1 

3, 3 

33 3> ^ 

e stupas' ? , 

O f" ITf ^Ti ft Q 

H U U L*l3t*3 


j ' 

3, 3' 


31 33 a 3> 

'stupa* j 

r stupa j ) 

A H 


33 22 

33 j; 3 jt 

bu<Ji 3? 


7 ,? 


? / 


33 33 * 33 



5 *S 



)} 33 2 j, 

*varana* ,, 


* * 

Sat. >j 







JJ JJ JJ 33 
33 33 * >J 

toddy) f ( s talf - 

toddy- ('tall'-} 


33 4 

1 <l 41 ^ 3) 

fibre ^or 3J 

fibre- or 


1 VJ 



33 3? * ' 

& O 97 

Kaus j? 





jj J. 

,, 6 

J J J? '' 
33 3J 3 

'involving 3 , 




j? w 


c tuda f 




" ^ 

!! s 

*upastarana j aj 


ft * 


3 J 

,, text 

e asand!'s' >, 

f -+ Mb * 

y 7 
* 5 


3, l^ 

S J> " 

V * 







of n. 6 


'uruni* re 







3) text 






? , 


33 n- 6 









33 3? 33 









, 7 








J 3 33 >3 









J3 33 " 









33 J3 ^ 









3' 33 3) 








33 33 ^ 










i J 

'khaddar? 5 


'khaddar' ? 





J3 33 2 









,, text 









n. 3 











1 3 3 ? 33 

* j 






33 33 33 









31 33 **- 









33 33 5 


"Rak. " 







33 33 ^ 









33 33 10 




V. I. 





, ? text 


'opasa.' 2 . 







33 n. 2 

3 * 








33 33 ^ 




V. I. 





33 3 1 * *- 

^ j 








... 4 











Bra. Siit. and 


Bra. and Silt 






33 33 









33 33 ^ 


'vidava 3 








; i 

'parivrkti 1 







n. 3 










i) 33 ^ 









33 33 ^* 









33 33 ** 









,, -, 11 

/ i^ 








3, 33 2 









33 33 ' 

7 J 








33 33 $ 


11, 15, 


11. 15. 





33 33 33 


f jani' 







33 33 " 









>} 3? J- 

3 ^ 

















33 33 

? , 








n- 5 

7- 1 








33 33 ** 




etc.) ; 

? 3 




33 33 10 









33 33 3J 









33 33 5 









33 33 J- 









,, text 









n. 3 









33 33 8 


VahyakascO pa 

VahyakaHc Opa 

j^H^k *' J 



r^- vahyakah. 





3>J 33 ;, 




" (tesam 





> 33 ^ 








Page 138, line 14 

?J 141 ,3 9 
,, 148 3 














156 26 
162 I 


164, line 29 

171, line 4 

33 33 33 3J 

3, 172 4 

. 174 6 

181 10 

183 3 

184 2 


}j J.^^- tj 

33 33 























> > 











































J J 








3 J 







































n. 4 for 

n, 7 







33 }} ^ 33 

headline, }) 

of text, should 
,, n. 2 fof 

3? 3? 33 

33 33 ^ 33 

, , text 

3.1 33 

3J 33 

33 33 

.'3 J 3 


of n. 






n. 1 
headline , 



n. 2 

















33 3 

n. 2 

33 4 




n. 5 

33 * 

n. 9 




33 n. 2 













'svasutayam' read 

op. cit. 


"niahavsini va 

4 Of. 


be line 1 of text, 
31 n. read 

5 u : 3 , 
25. 0. 


op. cit., 

"maharsin iva 
4 Of. 

1 IV. 

page 165* 


11 j 




tliemselves (e.g. 
1 Mbh. 










7, 45-8, 51 





"apsaras* like 






should be line 22 
3, 23 









Bhanunati ,$ 

Laksmina f , 

Vabhuva Karani 

Suta 3 , 

Ka Siit. 

Paiicala }3 





11 3 









themselves; e.g. 
1 Mbh. 



7 3 45-8, 51 





16, 9 



of text, 


33 > 



vabniiva karanl- 

KS. But. 
132, 105